AuroraMiddleware/glslang
1
0
Fork 0
mirror of https://github.com/KhronosGroup/glslang synced 2024-11-08 19:40:06 +00:00
Code Issues Packages Projects Releases Wiki Activity
vulkan-sdk-1.3.280
glslang/Test/baseResults/hlsl.intrinsics.vert.out

4198 lines
195 KiB
Plaintext
Raw Permalink Normal View History

Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
hlsl.intrinsics.vert
HLSL: Emit the OpSource HLSL instruction for HLSL, using new headers.
2017-04-07 21:33:08 +00:00
Shader version: 500
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:2 Function Definition: VertexShaderFunctionS(f1;f1;f1;u1;u1; ( temp float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:2 Function Parameters:
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:2 'inF0' ( in float)
0:2 'inF1' ( in float)
0:2 'inF2' ( in float)
0:2 'inU0' ( in uint)
0:2 'inU1' ( in uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:3 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:3 Convert float to bool ( temp bool)
0:3 'inF0' ( in float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:4 Absolute value ( temp float)
0:4 'inF0' ( in float)
0:5 arc cosine ( temp float)
0:5 'inF0' ( in float)
0:6 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:6 Convert float to bool ( temp bool)
0:6 'inF0' ( in float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:7 arc sine ( temp float)
0:7 'inF0' ( in float)
0:8 floatBitsToInt ( temp int)
0:8 'inF0' ( in float)
0:9 floatBitsToUint ( temp uint)
0:9 'inF0' ( in float)
0:10 intBitsToFloat ( temp float)
0:10 'inU0' ( in uint)
0:12 arc tangent ( temp float)
0:12 'inF0' ( in float)
0:13 arc tangent ( temp float)
0:13 'inF0' ( in float)
0:13 'inF1' ( in float)
0:14 Ceiling ( temp float)
0:14 'inF0' ( in float)
0:15 clamp ( temp float)
0:15 'inF0' ( in float)
0:15 'inF1' ( in float)
0:15 'inF2' ( in float)
0:16 cosine ( temp float)
0:16 'inF0' ( in float)
0:17 hyp. cosine ( temp float)
0:17 'inF0' ( in float)
0:18 bitCount ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:18 Constant:
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
0:18 7 (const int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:19 degrees ( temp float)
0:19 'inF0' ( in float)
0:23 exp ( temp float)
0:23 'inF0' ( in float)
0:24 exp2 ( temp float)
0:24 'inF0' ( in float)
0:25 findMSB ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:25 Constant:
0:25 7 (const int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:26 findLSB ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:26 Constant:
0:26 7 (const int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:27 Floor ( temp float)
0:27 'inF0' ( in float)
0:29 mod ( temp float)
0:29 'inF0' ( in float)
0:29 'inF1' ( in float)
0:30 Fraction ( temp float)
0:30 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:31 isinf ( temp bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:31 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:32 isnan ( temp bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:32 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:33 ldexp ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:33 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:33 'inF1' ( in float)
0:34 mix ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:34 'inF0' ( in float)
0:34 'inF1' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:34 'inF2' ( in float)
0:35 log ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:35 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:36 component-wise multiply ( temp float)
0:36 log2 ( temp float)
0:36 'inF0' ( in float)
0:36 Constant:
0:36 0.301030
0:37 log2 ( temp float)
0:37 'inF0' ( in float)
0:38 max ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:38 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:38 'inF1' ( in float)
0:39 min ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:39 'inF0' ( in float)
0:39 'inF1' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:41 pow ( temp float)
0:41 'inF0' ( in float)
0:41 'inF1' ( in float)
0:42 radians ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:42 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:43 bitFieldReverse ( temp int)
0:43 Constant:
0:43 2 (const int)
0:44 roundEven ( temp float)
0:44 'inF0' ( in float)
0:45 inverse sqrt ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:45 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:46 clamp ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:46 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:46 Constant:
0:46 0.000000
0:46 Constant:
0:46 1.000000
0:47 Sign ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:47 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:48 sine ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:48 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:49 Sequence
0:49 move second child to first child ( temp float)
0:49 'inF1' ( in float)
0:49 sine ( temp float)
0:49 'inF0' ( in float)
0:49 move second child to first child ( temp float)
0:49 'inF2' ( in float)
0:49 cosine ( temp float)
0:49 'inF0' ( in float)
0:50 hyp. sine ( temp float)
0:50 'inF0' ( in float)
0:51 smoothstep ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:51 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:51 'inF1' ( in float)
0:51 'inF2' ( in float)
0:52 sqrt ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:52 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:53 step ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:53 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:53 'inF1' ( in float)
0:54 tangent ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:54 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:55 hyp. tangent ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:55 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:57 trunc ( temp float)
0:57 'inF0' ( in float)
0:59 Branch: Return with expression
0:59 Constant:
0:59 0.000000
0:63 Function Definition: VertexShaderFunction1(vf1;vf1;vf1; ( temp 1-component vector of float)
0:63 Function Parameters:
0:63 'inF0' ( in 1-component vector of float)
0:63 'inF1' ( in 1-component vector of float)
0:63 'inF2' ( in 1-component vector of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:65 Branch: Return with expression
0:65 Constant:
0:65 0.000000
0:69 Function Definition: VertexShaderFunction2(vf2;vf2;vf2;vu2;vu2; ( temp 2-component vector of float)
0:69 Function Parameters:
0:69 'inF0' ( in 2-component vector of float)
0:69 'inF1' ( in 2-component vector of float)
0:69 'inF2' ( in 2-component vector of float)
0:69 'inU0' ( in 2-component vector of uint)
0:69 'inU1' ( in 2-component vector of uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:70 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:70 Convert float to bool ( temp 2-component vector of bool)
0:70 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:71 Absolute value ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:71 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:72 arc cosine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:72 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:73 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:73 Convert float to bool ( temp 2-component vector of bool)
0:73 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:74 arc sine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:74 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:75 floatBitsToInt ( temp 2-component vector of int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:75 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:76 floatBitsToUint ( temp 2-component vector of uint)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:76 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:77 intBitsToFloat ( temp 2-component vector of float)
0:77 'inU0' ( in 2-component vector of uint)
0:79 arc tangent ( temp 2-component vector of float)
0:79 'inF0' ( in 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:80 arc tangent ( temp 2-component vector of float)
0:80 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:80 'inF1' ( in 2-component vector of float)
0:81 Ceiling ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:81 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:82 clamp ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:82 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:82 'inF1' ( in 2-component vector of float)
0:82 'inF2' ( in 2-component vector of float)
0:83 cosine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:83 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:84 hyp. cosine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:84 'inF0' ( in 2-component vector of float)
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:85 bitCount ( temp 2-component vector of int)
0:85 Constant:
0:85 7 (const int)
0:85 3 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:86 degrees ( temp 2-component vector of float)
0:86 'inF0' ( in 2-component vector of float)
0:87 distance ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:87 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:87 'inF1' ( in 2-component vector of float)
0:88 dot-product ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:88 'inF0' ( in 2-component vector of float)
0:88 'inF1' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:92 exp ( temp 2-component vector of float)
0:92 'inF0' ( in 2-component vector of float)
0:93 exp2 ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:93 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:94 face-forward ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:94 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:94 'inF1' ( in 2-component vector of float)
0:94 'inF2' ( in 2-component vector of float)
0:95 findMSB ( temp int)
0:95 Constant:
0:95 7 (const int)
0:96 findLSB ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:96 Constant:
0:96 7 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:97 Floor ( temp 2-component vector of float)
0:97 'inF0' ( in 2-component vector of float)
0:99 mod ( temp 2-component vector of float)
0:99 'inF0' ( in 2-component vector of float)
0:99 'inF1' ( in 2-component vector of float)
0:100 Fraction ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:100 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:101 isinf ( temp 2-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:101 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:102 isnan ( temp 2-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:102 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:103 ldexp ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:103 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:103 'inF1' ( in 2-component vector of float)
0:104 mix ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:104 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:104 'inF1' ( in 2-component vector of float)
0:104 'inF2' ( in 2-component vector of float)
0:105 length ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:105 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:106 log ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:106 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:107 vector-scale ( temp 2-component vector of float)
0:107 log2 ( temp 2-component vector of float)
0:107 'inF0' ( in 2-component vector of float)
0:107 Constant:
0:107 0.301030
0:108 log2 ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:108 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:109 max ( temp 2-component vector of float)
0:109 'inF0' ( in 2-component vector of float)
0:109 'inF1' ( in 2-component vector of float)
0:110 min ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:110 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:110 'inF1' ( in 2-component vector of float)
0:112 normalize ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:112 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:113 pow ( temp 2-component vector of float)
0:113 'inF0' ( in 2-component vector of float)
0:113 'inF1' ( in 2-component vector of float)
0:114 radians ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:114 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:115 reflect ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:115 'inF0' ( in 2-component vector of float)
0:115 'inF1' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:116 refract ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:116 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:116 'inF1' ( in 2-component vector of float)
0:116 Constant:
0:116 2.000000
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:117 bitFieldReverse ( temp 2-component vector of int)
0:117 Constant:
0:117 1 (const int)
0:117 2 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:118 roundEven ( temp 2-component vector of float)
0:118 'inF0' ( in 2-component vector of float)
0:119 inverse sqrt ( temp 2-component vector of float)
0:119 'inF0' ( in 2-component vector of float)
0:120 clamp ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:120 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:120 Constant:
0:120 0.000000
0:120 Constant:
0:120 1.000000
0:121 Sign ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:121 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:122 sine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:122 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:123 Sequence
0:123 move second child to first child ( temp 2-component vector of float)
0:123 'inF1' ( in 2-component vector of float)
0:123 sine ( temp 2-component vector of float)
0:123 'inF0' ( in 2-component vector of float)
0:123 move second child to first child ( temp 2-component vector of float)
0:123 'inF2' ( in 2-component vector of float)
0:123 cosine ( temp 2-component vector of float)
0:123 'inF0' ( in 2-component vector of float)
0:124 hyp. sine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:124 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:125 smoothstep ( temp 2-component vector of float)
0:125 'inF0' ( in 2-component vector of float)
0:125 'inF1' ( in 2-component vector of float)
0:125 'inF2' ( in 2-component vector of float)
0:126 sqrt ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:126 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:127 step ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:127 'inF0' ( in 2-component vector of float)
0:127 'inF1' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:128 tangent ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:128 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:129 hyp. tangent ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:129 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:131 trunc ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:131 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:134 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:134 Constant:
0:134 1.000000
0:134 2.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:138 Function Definition: VertexShaderFunction3(vf3;vf3;vf3;vu3;vu3; ( temp 3-component vector of float)
0:138 Function Parameters:
0:138 'inF0' ( in 3-component vector of float)
0:138 'inF1' ( in 3-component vector of float)
0:138 'inF2' ( in 3-component vector of float)
0:138 'inU0' ( in 3-component vector of uint)
0:138 'inU1' ( in 3-component vector of uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:139 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:139 Convert float to bool ( temp 3-component vector of bool)
0:139 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:140 Absolute value ( temp 3-component vector of float)
0:140 'inF0' ( in 3-component vector of float)
0:141 arc cosine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:141 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:142 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:142 Convert float to bool ( temp 3-component vector of bool)
0:142 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:143 arc sine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:143 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:144 floatBitsToInt ( temp 3-component vector of int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:144 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:145 floatBitsToUint ( temp 3-component vector of uint)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:145 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:146 intBitsToFloat ( temp 3-component vector of float)
0:146 'inU0' ( in 3-component vector of uint)
0:148 arc tangent ( temp 3-component vector of float)
0:148 'inF0' ( in 3-component vector of float)
0:149 arc tangent ( temp 3-component vector of float)
0:149 'inF0' ( in 3-component vector of float)
0:149 'inF1' ( in 3-component vector of float)
0:150 Ceiling ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:150 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:151 clamp ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:151 'inF0' ( in 3-component vector of float)
0:151 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:151 'inF2' ( in 3-component vector of float)
0:152 cosine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:152 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:153 hyp. cosine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:153 'inF0' ( in 3-component vector of float)
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:154 bitCount ( temp 3-component vector of int)
0:154 Constant:
0:154 7 (const int)
0:154 3 (const int)
0:154 5 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:155 cross-product ( temp 3-component vector of float)
0:155 'inF0' ( in 3-component vector of float)
0:155 'inF1' ( in 3-component vector of float)
0:156 degrees ( temp 3-component vector of float)
0:156 'inF0' ( in 3-component vector of float)
0:157 distance ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:157 'inF0' ( in 3-component vector of float)
0:157 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:158 dot-product ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:158 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:158 'inF1' ( in 3-component vector of float)
0:162 exp ( temp 3-component vector of float)
0:162 'inF0' ( in 3-component vector of float)
0:163 exp2 ( temp 3-component vector of float)
0:163 'inF0' ( in 3-component vector of float)
0:164 face-forward ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:164 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:164 'inF1' ( in 3-component vector of float)
0:164 'inF2' ( in 3-component vector of float)
0:165 findMSB ( temp int)
0:165 Constant:
0:165 7 (const int)
0:166 findLSB ( temp int)
0:166 Constant:
0:166 7 (const int)
0:167 Floor ( temp 3-component vector of float)
0:167 'inF0' ( in 3-component vector of float)
0:169 mod ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:169 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:169 'inF1' ( in 3-component vector of float)
0:170 Fraction ( temp 3-component vector of float)
0:170 'inF0' ( in 3-component vector of float)
0:171 isinf ( temp 3-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:171 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:172 isnan ( temp 3-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:172 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:173 ldexp ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:173 'inF0' ( in 3-component vector of float)
0:173 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:174 mix ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:174 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:174 'inF1' ( in 3-component vector of float)
0:174 'inF2' ( in 3-component vector of float)
0:175 length ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:175 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:176 log ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:176 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:177 vector-scale ( temp 3-component vector of float)
0:177 log2 ( temp 3-component vector of float)
0:177 'inF0' ( in 3-component vector of float)
0:177 Constant:
0:177 0.301030
0:178 log2 ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:178 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:179 max ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:179 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:179 'inF1' ( in 3-component vector of float)
0:180 min ( temp 3-component vector of float)
0:180 'inF0' ( in 3-component vector of float)
0:180 'inF1' ( in 3-component vector of float)
0:182 normalize ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:182 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:183 pow ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:183 'inF0' ( in 3-component vector of float)
0:183 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:184 radians ( temp 3-component vector of float)
0:184 'inF0' ( in 3-component vector of float)
0:185 reflect ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:185 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:185 'inF1' ( in 3-component vector of float)
0:186 refract ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:186 'inF0' ( in 3-component vector of float)
0:186 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:186 Constant:
0:186 2.000000
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:187 bitFieldReverse ( temp 3-component vector of int)
0:187 Constant:
0:187 1 (const int)
0:187 2 (const int)
0:187 3 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:188 roundEven ( temp 3-component vector of float)
0:188 'inF0' ( in 3-component vector of float)
0:189 inverse sqrt ( temp 3-component vector of float)
0:189 'inF0' ( in 3-component vector of float)
0:190 clamp ( temp 3-component vector of float)
0:190 'inF0' ( in 3-component vector of float)
0:190 Constant:
0:190 0.000000
0:190 Constant:
0:190 1.000000
0:191 Sign ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:191 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:192 sine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:192 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:193 Sequence
0:193 move second child to first child ( temp 3-component vector of float)
0:193 'inF1' ( in 3-component vector of float)
0:193 sine ( temp 3-component vector of float)
0:193 'inF0' ( in 3-component vector of float)
0:193 move second child to first child ( temp 3-component vector of float)
0:193 'inF2' ( in 3-component vector of float)
0:193 cosine ( temp 3-component vector of float)
0:193 'inF0' ( in 3-component vector of float)
0:194 hyp. sine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:194 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:195 smoothstep ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:195 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:195 'inF1' ( in 3-component vector of float)
0:195 'inF2' ( in 3-component vector of float)
0:196 sqrt ( temp 3-component vector of float)
0:196 'inF0' ( in 3-component vector of float)
0:197 step ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:197 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:197 'inF1' ( in 3-component vector of float)
0:198 tangent ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:198 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:199 hyp. tangent ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:199 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:201 trunc ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:201 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:204 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:204 Constant:
0:204 1.000000
0:204 2.000000
0:204 3.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:208 Function Definition: VertexShaderFunction4(vf4;vf4;vf4;vu4;vu4; ( temp 4-component vector of float)
0:208 Function Parameters:
0:208 'inF0' ( in 4-component vector of float)
0:208 'inF1' ( in 4-component vector of float)
0:208 'inF2' ( in 4-component vector of float)
0:208 'inU0' ( in 4-component vector of uint)
0:208 'inU1' ( in 4-component vector of uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:209 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:209 Convert float to bool ( temp 4-component vector of bool)
0:209 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:210 Absolute value ( temp 4-component vector of float)
0:210 'inF0' ( in 4-component vector of float)
0:211 arc cosine ( temp 4-component vector of float)
0:211 'inF0' ( in 4-component vector of float)
0:212 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:212 Convert float to bool ( temp 4-component vector of bool)
0:212 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:213 arc sine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:213 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:214 floatBitsToInt ( temp 4-component vector of int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:214 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:215 floatBitsToUint ( temp 4-component vector of uint)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:215 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:216 intBitsToFloat ( temp 4-component vector of float)
0:216 'inU0' ( in 4-component vector of uint)
0:218 arc tangent ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:218 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:219 arc tangent ( temp 4-component vector of float)
0:219 'inF0' ( in 4-component vector of float)
0:219 'inF1' ( in 4-component vector of float)
0:220 Ceiling ( temp 4-component vector of float)
0:220 'inF0' ( in 4-component vector of float)
0:221 clamp ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:221 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:221 'inF1' ( in 4-component vector of float)
0:221 'inF2' ( in 4-component vector of float)
0:222 cosine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:222 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:223 hyp. cosine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:223 'inF0' ( in 4-component vector of float)
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:224 bitCount ( temp 4-component vector of int)
0:224 Constant:
0:224 7 (const int)
0:224 3 (const int)
0:224 5 (const int)
0:224 2 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:225 degrees ( temp 4-component vector of float)
0:225 'inF0' ( in 4-component vector of float)
0:226 distance ( temp float)
0:226 'inF0' ( in 4-component vector of float)
0:226 'inF1' ( in 4-component vector of float)
0:227 dot-product ( temp float)
0:227 'inF0' ( in 4-component vector of float)
0:227 'inF1' ( in 4-component vector of float)
0:228 Construct vec4 ( temp 4-component vector of float)
0:228 Constant:
0:228 1.000000
0:228 component-wise multiply ( temp float)
0:228 direct index ( temp float)
0:228 'inF0' ( in 4-component vector of float)
0:228 Constant:
0:228 1 (const int)
0:228 direct index ( temp float)
0:228 'inF1' ( in 4-component vector of float)
0:228 Constant:
0:228 1 (const int)
0:228 direct index ( temp float)
0:228 'inF0' ( in 4-component vector of float)
0:228 Constant:
0:228 2 (const int)
0:228 direct index ( temp float)
0:228 'inF1' ( in 4-component vector of float)
0:228 Constant:
0:228 3 (const int)
0:232 exp ( temp 4-component vector of float)
0:232 'inF0' ( in 4-component vector of float)
0:233 exp2 ( temp 4-component vector of float)
0:233 'inF0' ( in 4-component vector of float)
0:234 face-forward ( temp 4-component vector of float)
0:234 'inF0' ( in 4-component vector of float)
0:234 'inF1' ( in 4-component vector of float)
0:234 'inF2' ( in 4-component vector of float)
0:235 findMSB ( temp int)
0:235 Constant:
0:235 7 (const int)
0:236 findLSB ( temp int)
0:236 Constant:
0:236 7 (const int)
0:237 Floor ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:237 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:239 mod ( temp 4-component vector of float)
0:239 'inF0' ( in 4-component vector of float)
0:239 'inF1' ( in 4-component vector of float)
0:240 Fraction ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:240 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:241 isinf ( temp 4-component vector of bool)
0:241 'inF0' ( in 4-component vector of float)
0:242 isnan ( temp 4-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:242 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:243 ldexp ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:243 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:243 'inF1' ( in 4-component vector of float)
0:244 mix ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:244 'inF0' ( in 4-component vector of float)
0:244 'inF1' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:244 'inF2' ( in 4-component vector of float)
0:245 length ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:245 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:246 log ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:246 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:247 vector-scale ( temp 4-component vector of float)
0:247 log2 ( temp 4-component vector of float)
0:247 'inF0' ( in 4-component vector of float)
0:247 Constant:
0:247 0.301030
0:248 log2 ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:248 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:249 max ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:249 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:249 'inF1' ( in 4-component vector of float)
0:250 min ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:250 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:250 'inF1' ( in 4-component vector of float)
0:252 normalize ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:252 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:253 pow ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:253 'inF0' ( in 4-component vector of float)
0:253 'inF1' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:254 radians ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:254 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:255 reflect ( temp 4-component vector of float)
0:255 'inF0' ( in 4-component vector of float)
0:255 'inF1' ( in 4-component vector of float)
0:256 refract ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:256 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:256 'inF1' ( in 4-component vector of float)
0:256 Constant:
0:256 2.000000
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:257 bitFieldReverse ( temp 4-component vector of int)
0:257 Constant:
0:257 1 (const int)
0:257 2 (const int)
0:257 3 (const int)
0:257 4 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:258 roundEven ( temp 4-component vector of float)
0:258 'inF0' ( in 4-component vector of float)
0:259 inverse sqrt ( temp 4-component vector of float)
0:259 'inF0' ( in 4-component vector of float)
0:260 clamp ( temp 4-component vector of float)
0:260 'inF0' ( in 4-component vector of float)
0:260 Constant:
0:260 0.000000
0:260 Constant:
0:260 1.000000
0:261 Sign ( temp 4-component vector of float)
0:261 'inF0' ( in 4-component vector of float)
0:262 sine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:262 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:263 Sequence
0:263 move second child to first child ( temp 4-component vector of float)
0:263 'inF1' ( in 4-component vector of float)
0:263 sine ( temp 4-component vector of float)
0:263 'inF0' ( in 4-component vector of float)
0:263 move second child to first child ( temp 4-component vector of float)
0:263 'inF2' ( in 4-component vector of float)
0:263 cosine ( temp 4-component vector of float)
0:263 'inF0' ( in 4-component vector of float)
0:264 hyp. sine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:264 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:265 smoothstep ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:265 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:265 'inF1' ( in 4-component vector of float)
0:265 'inF2' ( in 4-component vector of float)
0:266 sqrt ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:266 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:267 step ( temp 4-component vector of float)
0:267 'inF0' ( in 4-component vector of float)
0:267 'inF1' ( in 4-component vector of float)
0:268 tangent ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:268 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:269 hyp. tangent ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:269 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:271 trunc ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:271 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:274 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:274 Constant:
0:274 1.000000
0:274 2.000000
0:274 3.000000
0:274 4.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:331 Function Definition: VertexShaderFunction2x2(mf22;mf22;mf22; ( temp 2X2 matrix of float)
0:331 Function Parameters:
0:331 'inF0' ( in 2X2 matrix of float)
0:331 'inF1' ( in 2X2 matrix of float)
0:331 'inF2' ( in 2X2 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:333 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:333 Convert float to bool ( temp 2X2 matrix of bool)
0:333 'inF0' ( in 2X2 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:333 Absolute value ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 arc cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:333 Convert float to bool ( temp 2X2 matrix of bool)
0:333 'inF0' ( in 2X2 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:333 arc sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 arc tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 arc tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 Ceiling ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 clamp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 hyp. cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 degrees ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 determinant ( temp float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 exp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 exp2 ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 findMSB ( temp int)
0:333 Constant:
0:333 7 (const int)
0:333 findLSB ( temp int)
0:333 Constant:
0:333 7 (const int)
0:333 Floor ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 mod ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 Fraction ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 ldexp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 mix ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 log ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 matrix-scale ( temp 2X2 matrix of float)
0:333 log2 ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 Constant:
0:333 0.301030
0:333 log2 ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 max ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 min ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 pow ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 radians ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 roundEven ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 inverse sqrt ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 clamp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 Constant:
0:333 0.000000
0:333 Constant:
0:333 1.000000
0:333 Sign ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 Sequence
0:333 move second child to first child ( temp 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 move second child to first child ( temp 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 hyp. sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 smoothstep ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 sqrt ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 step ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 hyp. tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 transpose ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 trunc ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:336 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:336 Constant:
0:336 2.000000
0:336 2.000000
0:336 2.000000
0:336 2.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:340 Function Definition: VertexShaderFunction3x3(mf33;mf33;mf33; ( temp 3X3 matrix of float)
0:340 Function Parameters:
0:340 'inF0' ( in 3X3 matrix of float)
0:340 'inF1' ( in 3X3 matrix of float)
0:340 'inF2' ( in 3X3 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:342 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:342 Convert float to bool ( temp 3X3 matrix of bool)
0:342 'inF0' ( in 3X3 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:342 Absolute value ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 arc cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:342 Convert float to bool ( temp 3X3 matrix of bool)
0:342 'inF0' ( in 3X3 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:342 arc sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 arc tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 arc tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 Ceiling ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 clamp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 hyp. cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 degrees ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 determinant ( temp float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 exp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 exp2 ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 findMSB ( temp int)
0:342 Constant:
0:342 7 (const int)
0:342 findLSB ( temp int)
0:342 Constant:
0:342 7 (const int)
0:342 Floor ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 mod ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 Fraction ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 ldexp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 mix ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 log ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 matrix-scale ( temp 3X3 matrix of float)
0:342 log2 ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 Constant:
0:342 0.301030
0:342 log2 ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 max ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 min ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 pow ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 radians ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 roundEven ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 inverse sqrt ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 clamp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 Constant:
0:342 0.000000
0:342 Constant:
0:342 1.000000
0:342 Sign ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 Sequence
0:342 move second child to first child ( temp 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 move second child to first child ( temp 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 hyp. sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 smoothstep ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 sqrt ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 step ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 hyp. tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 transpose ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 trunc ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:345 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:345 Constant:
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:349 Function Definition: VertexShaderFunction4x4(mf44;mf44;mf44; ( temp 4X4 matrix of float)
0:349 Function Parameters:
0:349 'inF0' ( in 4X4 matrix of float)
0:349 'inF1' ( in 4X4 matrix of float)
0:349 'inF2' ( in 4X4 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:351 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:351 Convert float to bool ( temp 4X4 matrix of bool)
0:351 'inF0' ( in 4X4 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:351 Absolute value ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 arc cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:351 Convert float to bool ( temp 4X4 matrix of bool)
0:351 'inF0' ( in 4X4 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:351 arc sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 arc tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 arc tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 Ceiling ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 clamp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 hyp. cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 degrees ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 determinant ( temp float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 exp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 exp2 ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 findMSB ( temp int)
0:351 Constant:
0:351 7 (const int)
0:351 findLSB ( temp int)
0:351 Constant:
0:351 7 (const int)
0:351 Floor ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 mod ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 Fraction ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 ldexp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 mix ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 log ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 matrix-scale ( temp 4X4 matrix of float)
0:351 log2 ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 Constant:
0:351 0.301030
0:351 log2 ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 max ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 min ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 pow ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 radians ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 roundEven ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 inverse sqrt ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 clamp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 Constant:
0:351 0.000000
0:351 Constant:
0:351 1.000000
0:351 Sign ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 Sequence
0:351 move second child to first child ( temp 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 move second child to first child ( temp 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 hyp. sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 smoothstep ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 sqrt ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 step ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 hyp. tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 transpose ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 trunc ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:354 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:354 Constant:
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:372 Function Definition: TestGenMul2(f1;f1;vf2;vf2;mf22;mf22; ( temp void)
0:372 Function Parameters:
0:372 'inF0' ( in float)
0:372 'inF1' ( in float)
0:372 'inFV0' ( in 2-component vector of float)
0:372 'inFV1' ( in 2-component vector of float)
0:372 'inFM0' ( in 2X2 matrix of float)
0:372 'inFM1' ( in 2X2 matrix of float)
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:373 Sequence
0:373 move second child to first child ( temp float)
0:373 'r0' ( temp float)
0:373 component-wise multiply ( temp float)
0:373 'inF1' ( in float)
0:373 'inF0' ( in float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r1' ( temp 2-component vector of float)
0:373 vector-scale ( temp 2-component vector of float)
0:373 'inF0' ( in float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r2' ( temp 2-component vector of float)
0:373 vector-scale ( temp 2-component vector of float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 'inF0' ( in float)
0:373 Sequence
0:373 move second child to first child ( temp float)
0:373 'r3' ( temp float)
0:373 dot-product ( temp float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 'inFV1' ( in 2-component vector of float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r4' ( temp 2-component vector of float)
0:373 vector-times-matrix ( temp 2-component vector of float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r5' ( temp 2-component vector of float)
0:373 matrix-times-vector ( temp 2-component vector of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 Sequence
0:373 move second child to first child ( temp 2X2 matrix of float)
0:373 'r6' ( temp 2X2 matrix of float)
0:373 matrix-scale ( temp 2X2 matrix of float)
0:373 'inF0' ( in float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 Sequence
0:373 move second child to first child ( temp 2X2 matrix of float)
0:373 'r7' ( temp 2X2 matrix of float)
0:373 matrix-scale ( temp 2X2 matrix of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 'inF0' ( in float)
0:373 Sequence
0:373 move second child to first child ( temp 2X2 matrix of float)
0:373 'r8' ( temp 2X2 matrix of float)
0:373 matrix-multiply ( temp 2X2 matrix of float)
0:373 'inFM1' ( in 2X2 matrix of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:379 Function Definition: TestGenMul3(f1;f1;vf3;vf3;mf33;mf33; ( temp void)
0:379 Function Parameters:
0:379 'inF0' ( in float)
0:379 'inF1' ( in float)
0:379 'inFV0' ( in 3-component vector of float)
0:379 'inFV1' ( in 3-component vector of float)
0:379 'inFM0' ( in 3X3 matrix of float)
0:379 'inFM1' ( in 3X3 matrix of float)
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:380 Sequence
0:380 move second child to first child ( temp float)
0:380 'r0' ( temp float)
0:380 component-wise multiply ( temp float)
0:380 'inF1' ( in float)
0:380 'inF0' ( in float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r1' ( temp 3-component vector of float)
0:380 vector-scale ( temp 3-component vector of float)
0:380 'inF0' ( in float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r2' ( temp 3-component vector of float)
0:380 vector-scale ( temp 3-component vector of float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 'inF0' ( in float)
0:380 Sequence
0:380 move second child to first child ( temp float)
0:380 'r3' ( temp float)
0:380 dot-product ( temp float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 'inFV1' ( in 3-component vector of float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r4' ( temp 3-component vector of float)
0:380 vector-times-matrix ( temp 3-component vector of float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r5' ( temp 3-component vector of float)
0:380 matrix-times-vector ( temp 3-component vector of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 Sequence
0:380 move second child to first child ( temp 3X3 matrix of float)
0:380 'r6' ( temp 3X3 matrix of float)
0:380 matrix-scale ( temp 3X3 matrix of float)
0:380 'inF0' ( in float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 Sequence
0:380 move second child to first child ( temp 3X3 matrix of float)
0:380 'r7' ( temp 3X3 matrix of float)
0:380 matrix-scale ( temp 3X3 matrix of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 'inF0' ( in float)
0:380 Sequence
0:380 move second child to first child ( temp 3X3 matrix of float)
0:380 'r8' ( temp 3X3 matrix of float)
0:380 matrix-multiply ( temp 3X3 matrix of float)
0:380 'inFM1' ( in 3X3 matrix of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:386 Function Definition: TestGenMul4(f1;f1;vf4;vf4;mf44;mf44; ( temp void)
0:386 Function Parameters:
0:386 'inF0' ( in float)
0:386 'inF1' ( in float)
0:386 'inFV0' ( in 4-component vector of float)
0:386 'inFV1' ( in 4-component vector of float)
0:386 'inFM0' ( in 4X4 matrix of float)
0:386 'inFM1' ( in 4X4 matrix of float)
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:387 Sequence
0:387 move second child to first child ( temp float)
0:387 'r0' ( temp float)
0:387 component-wise multiply ( temp float)
0:387 'inF1' ( in float)
0:387 'inF0' ( in float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r1' ( temp 4-component vector of float)
0:387 vector-scale ( temp 4-component vector of float)
0:387 'inF0' ( in float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r2' ( temp 4-component vector of float)
0:387 vector-scale ( temp 4-component vector of float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 'inF0' ( in float)
0:387 Sequence
0:387 move second child to first child ( temp float)
0:387 'r3' ( temp float)
0:387 dot-product ( temp float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 'inFV1' ( in 4-component vector of float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r4' ( temp 4-component vector of float)
0:387 vector-times-matrix ( temp 4-component vector of float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r5' ( temp 4-component vector of float)
0:387 matrix-times-vector ( temp 4-component vector of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 Sequence
0:387 move second child to first child ( temp 4X4 matrix of float)
0:387 'r6' ( temp 4X4 matrix of float)
0:387 matrix-scale ( temp 4X4 matrix of float)
0:387 'inF0' ( in float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 Sequence
0:387 move second child to first child ( temp 4X4 matrix of float)
0:387 'r7' ( temp 4X4 matrix of float)
0:387 matrix-scale ( temp 4X4 matrix of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 'inF0' ( in float)
0:387 Sequence
0:387 move second child to first child ( temp 4X4 matrix of float)
0:387 'r8' ( temp 4X4 matrix of float)
0:387 matrix-multiply ( temp 4X4 matrix of float)
0:387 'inFM1' ( in 4X4 matrix of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:396 Function Definition: TestGenMulNxM(f1;f1;vf2;vf3;mf23;mf32;mf33;mf34;mf24; ( temp void)
0:396 Function Parameters:
0:396 'inF0' ( in float)
0:396 'inF1' ( in float)
0:396 'inFV2' ( in 2-component vector of float)
0:396 'inFV3' ( in 3-component vector of float)
0:396 'inFM2x3' ( in 2X3 matrix of float)
0:396 'inFM3x2' ( in 3X2 matrix of float)
0:396 'inFM3x3' ( in 3X3 matrix of float)
0:396 'inFM3x4' ( in 3X4 matrix of float)
0:396 'inFM2x4' ( in 2X4 matrix of float)
Add lerp, fix sincos return type, ret type tests, non-square mats, HLSL->AST tx semantics
2016-06-20 19:26:59 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:397 Sequence
0:397 move second child to first child ( temp float)
0:397 'r00' ( temp float)
0:397 component-wise multiply ( temp float)
0:397 'inF1' ( in float)
0:397 'inF0' ( in float)
0:398 Sequence
0:398 move second child to first child ( temp 2-component vector of float)
0:398 'r01' ( temp 2-component vector of float)
0:398 vector-scale ( temp 2-component vector of float)
0:398 'inF0' ( in float)
0:398 'inFV2' ( in 2-component vector of float)
0:399 Sequence
0:399 move second child to first child ( temp 3-component vector of float)
0:399 'r02' ( temp 3-component vector of float)
0:399 vector-scale ( temp 3-component vector of float)
0:399 'inF0' ( in float)
0:399 'inFV3' ( in 3-component vector of float)
0:400 Sequence
0:400 move second child to first child ( temp 2-component vector of float)
0:400 'r03' ( temp 2-component vector of float)
0:400 vector-scale ( temp 2-component vector of float)
0:400 'inFV2' ( in 2-component vector of float)
0:400 'inF0' ( in float)
0:401 Sequence
0:401 move second child to first child ( temp 3-component vector of float)
0:401 'r04' ( temp 3-component vector of float)
0:401 vector-scale ( temp 3-component vector of float)
0:401 'inFV3' ( in 3-component vector of float)
0:401 'inF0' ( in float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:402 Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:402 move second child to first child ( temp float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:402 'r05' ( temp float)
0:402 dot-product ( temp float)
0:402 'inFV2' ( in 2-component vector of float)
0:402 'inFV2' ( in 2-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:403 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:403 move second child to first child ( temp float)
0:403 'r06' ( temp float)
0:403 dot-product ( temp float)
0:403 'inFV3' ( in 3-component vector of float)
0:403 'inFV3' ( in 3-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:404 Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:404 move second child to first child ( temp 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:404 'r07' ( temp 3-component vector of float)
0:404 matrix-times-vector ( temp 3-component vector of float)
0:404 'inFM2x3' ( in 2X3 matrix of float)
0:404 'inFV2' ( in 2-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:405 Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:405 move second child to first child ( temp 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:405 'r08' ( temp 2-component vector of float)
0:405 matrix-times-vector ( temp 2-component vector of float)
0:405 'inFM3x2' ( in 3X2 matrix of float)
0:405 'inFV3' ( in 3-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:406 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:406 move second child to first child ( temp 2-component vector of float)
0:406 'r09' ( temp 2-component vector of float)
0:406 vector-times-matrix ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:406 'inFV3' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:406 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:407 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:407 move second child to first child ( temp 3-component vector of float)
0:407 'r10' ( temp 3-component vector of float)
0:407 vector-times-matrix ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:407 'inFV2' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:407 'inFM3x2' ( in 3X2 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:408 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:408 move second child to first child ( temp 2X3 matrix of float)
0:408 'r11' ( temp 2X3 matrix of float)
0:408 matrix-scale ( temp 2X3 matrix of float)
0:408 'inF0' ( in float)
0:408 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:409 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:409 move second child to first child ( temp 3X2 matrix of float)
0:409 'r12' ( temp 3X2 matrix of float)
0:409 matrix-scale ( temp 3X2 matrix of float)
0:409 'inF0' ( in float)
0:409 'inFM3x2' ( in 3X2 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:410 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:410 move second child to first child ( temp 2X2 matrix of float)
0:410 'r13' ( temp 2X2 matrix of float)
0:410 matrix-multiply ( temp 2X2 matrix of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:410 'inFM3x2' ( in 3X2 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:410 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:411 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:411 move second child to first child ( temp 2X3 matrix of float)
0:411 'r14' ( temp 2X3 matrix of float)
0:411 matrix-multiply ( temp 2X3 matrix of float)
0:411 'inFM3x3' ( in 3X3 matrix of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:411 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:412 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:412 move second child to first child ( temp 2X4 matrix of float)
0:412 'r15' ( temp 2X4 matrix of float)
0:412 matrix-multiply ( temp 2X4 matrix of float)
0:412 'inFM3x4' ( in 3X4 matrix of float)
0:412 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:413 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:413 move second child to first child ( temp 3X4 matrix of float)
0:413 'r16' ( temp 3X4 matrix of float)
0:413 matrix-multiply ( temp 3X4 matrix of float)
0:413 'inFM2x4' ( in 2X4 matrix of float)
0:413 'inFM3x2' ( in 3X2 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Linker Objects
Linked vertex stage:
Warn on HLSL not finding entry point. Issue #588.
2016-11-26 20:45:18 +00:00
WARNING: Linking vertex stage: Entry point not found
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
HLSL: Emit the OpSource HLSL instruction for HLSL, using new headers.
2017-04-07 21:33:08 +00:00
Shader version: 500
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:2 Function Definition: VertexShaderFunctionS(f1;f1;f1;u1;u1; ( temp float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:2 Function Parameters:
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:2 'inF0' ( in float)
0:2 'inF1' ( in float)
0:2 'inF2' ( in float)
0:2 'inU0' ( in uint)
0:2 'inU1' ( in uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:3 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:3 Convert float to bool ( temp bool)
0:3 'inF0' ( in float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:4 Absolute value ( temp float)
0:4 'inF0' ( in float)
0:5 arc cosine ( temp float)
0:5 'inF0' ( in float)
0:6 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:6 Convert float to bool ( temp bool)
0:6 'inF0' ( in float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:7 arc sine ( temp float)
0:7 'inF0' ( in float)
0:8 floatBitsToInt ( temp int)
0:8 'inF0' ( in float)
0:9 floatBitsToUint ( temp uint)
0:9 'inF0' ( in float)
0:10 intBitsToFloat ( temp float)
0:10 'inU0' ( in uint)
0:12 arc tangent ( temp float)
0:12 'inF0' ( in float)
0:13 arc tangent ( temp float)
0:13 'inF0' ( in float)
0:13 'inF1' ( in float)
0:14 Ceiling ( temp float)
0:14 'inF0' ( in float)
0:15 clamp ( temp float)
0:15 'inF0' ( in float)
0:15 'inF1' ( in float)
0:15 'inF2' ( in float)
0:16 cosine ( temp float)
0:16 'inF0' ( in float)
0:17 hyp. cosine ( temp float)
0:17 'inF0' ( in float)
0:18 bitCount ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:18 Constant:
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
0:18 7 (const int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:19 degrees ( temp float)
0:19 'inF0' ( in float)
0:23 exp ( temp float)
0:23 'inF0' ( in float)
0:24 exp2 ( temp float)
0:24 'inF0' ( in float)
0:25 findMSB ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:25 Constant:
0:25 7 (const int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:26 findLSB ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:26 Constant:
0:26 7 (const int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:27 Floor ( temp float)
0:27 'inF0' ( in float)
0:29 mod ( temp float)
0:29 'inF0' ( in float)
0:29 'inF1' ( in float)
0:30 Fraction ( temp float)
0:30 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:31 isinf ( temp bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:31 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:32 isnan ( temp bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:32 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:33 ldexp ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:33 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:33 'inF1' ( in float)
0:34 mix ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:34 'inF0' ( in float)
0:34 'inF1' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:34 'inF2' ( in float)
0:35 log ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:35 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:36 component-wise multiply ( temp float)
0:36 log2 ( temp float)
0:36 'inF0' ( in float)
0:36 Constant:
0:36 0.301030
0:37 log2 ( temp float)
0:37 'inF0' ( in float)
0:38 max ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:38 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:38 'inF1' ( in float)
0:39 min ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:39 'inF0' ( in float)
0:39 'inF1' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:41 pow ( temp float)
0:41 'inF0' ( in float)
0:41 'inF1' ( in float)
0:42 radians ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:42 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:43 bitFieldReverse ( temp int)
0:43 Constant:
0:43 2 (const int)
0:44 roundEven ( temp float)
0:44 'inF0' ( in float)
0:45 inverse sqrt ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:45 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:46 clamp ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:46 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:46 Constant:
0:46 0.000000
0:46 Constant:
0:46 1.000000
0:47 Sign ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:47 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:48 sine ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:48 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:49 Sequence
0:49 move second child to first child ( temp float)
0:49 'inF1' ( in float)
0:49 sine ( temp float)
0:49 'inF0' ( in float)
0:49 move second child to first child ( temp float)
0:49 'inF2' ( in float)
0:49 cosine ( temp float)
0:49 'inF0' ( in float)
0:50 hyp. sine ( temp float)
0:50 'inF0' ( in float)
0:51 smoothstep ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:51 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:51 'inF1' ( in float)
0:51 'inF2' ( in float)
0:52 sqrt ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:52 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:53 step ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:53 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:53 'inF1' ( in float)
0:54 tangent ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:54 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:55 hyp. tangent ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:55 'inF0' ( in float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:57 trunc ( temp float)
0:57 'inF0' ( in float)
0:59 Branch: Return with expression
0:59 Constant:
0:59 0.000000
0:63 Function Definition: VertexShaderFunction1(vf1;vf1;vf1; ( temp 1-component vector of float)
0:63 Function Parameters:
0:63 'inF0' ( in 1-component vector of float)
0:63 'inF1' ( in 1-component vector of float)
0:63 'inF2' ( in 1-component vector of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:65 Branch: Return with expression
0:65 Constant:
0:65 0.000000
0:69 Function Definition: VertexShaderFunction2(vf2;vf2;vf2;vu2;vu2; ( temp 2-component vector of float)
0:69 Function Parameters:
0:69 'inF0' ( in 2-component vector of float)
0:69 'inF1' ( in 2-component vector of float)
0:69 'inF2' ( in 2-component vector of float)
0:69 'inU0' ( in 2-component vector of uint)
0:69 'inU1' ( in 2-component vector of uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:70 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:70 Convert float to bool ( temp 2-component vector of bool)
0:70 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:71 Absolute value ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:71 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:72 arc cosine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:72 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:73 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:73 Convert float to bool ( temp 2-component vector of bool)
0:73 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:74 arc sine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:74 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:75 floatBitsToInt ( temp 2-component vector of int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:75 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:76 floatBitsToUint ( temp 2-component vector of uint)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:76 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:77 intBitsToFloat ( temp 2-component vector of float)
0:77 'inU0' ( in 2-component vector of uint)
0:79 arc tangent ( temp 2-component vector of float)
0:79 'inF0' ( in 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:80 arc tangent ( temp 2-component vector of float)
0:80 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:80 'inF1' ( in 2-component vector of float)
0:81 Ceiling ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:81 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:82 clamp ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:82 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:82 'inF1' ( in 2-component vector of float)
0:82 'inF2' ( in 2-component vector of float)
0:83 cosine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:83 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:84 hyp. cosine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:84 'inF0' ( in 2-component vector of float)
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:85 bitCount ( temp 2-component vector of int)
0:85 Constant:
0:85 7 (const int)
0:85 3 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:86 degrees ( temp 2-component vector of float)
0:86 'inF0' ( in 2-component vector of float)
0:87 distance ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:87 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:87 'inF1' ( in 2-component vector of float)
0:88 dot-product ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:88 'inF0' ( in 2-component vector of float)
0:88 'inF1' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:92 exp ( temp 2-component vector of float)
0:92 'inF0' ( in 2-component vector of float)
0:93 exp2 ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:93 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:94 face-forward ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:94 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:94 'inF1' ( in 2-component vector of float)
0:94 'inF2' ( in 2-component vector of float)
0:95 findMSB ( temp int)
0:95 Constant:
0:95 7 (const int)
0:96 findLSB ( temp int)
Implement atomic ops, bit conversions, fix fwidth stage mask, fix saturate dest modifier.
2016-06-13 15:22:28 +00:00
0:96 Constant:
0:96 7 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:97 Floor ( temp 2-component vector of float)
0:97 'inF0' ( in 2-component vector of float)
0:99 mod ( temp 2-component vector of float)
0:99 'inF0' ( in 2-component vector of float)
0:99 'inF1' ( in 2-component vector of float)
0:100 Fraction ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:100 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:101 isinf ( temp 2-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:101 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:102 isnan ( temp 2-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:102 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:103 ldexp ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:103 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:103 'inF1' ( in 2-component vector of float)
0:104 mix ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:104 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:104 'inF1' ( in 2-component vector of float)
0:104 'inF2' ( in 2-component vector of float)
0:105 length ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:105 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:106 log ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:106 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:107 vector-scale ( temp 2-component vector of float)
0:107 log2 ( temp 2-component vector of float)
0:107 'inF0' ( in 2-component vector of float)
0:107 Constant:
0:107 0.301030
0:108 log2 ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:108 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:109 max ( temp 2-component vector of float)
0:109 'inF0' ( in 2-component vector of float)
0:109 'inF1' ( in 2-component vector of float)
0:110 min ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:110 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:110 'inF1' ( in 2-component vector of float)
0:112 normalize ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:112 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:113 pow ( temp 2-component vector of float)
0:113 'inF0' ( in 2-component vector of float)
0:113 'inF1' ( in 2-component vector of float)
0:114 radians ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:114 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:115 reflect ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:115 'inF0' ( in 2-component vector of float)
0:115 'inF1' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:116 refract ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:116 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:116 'inF1' ( in 2-component vector of float)
0:116 Constant:
0:116 2.000000
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:117 bitFieldReverse ( temp 2-component vector of int)
0:117 Constant:
0:117 1 (const int)
0:117 2 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:118 roundEven ( temp 2-component vector of float)
0:118 'inF0' ( in 2-component vector of float)
0:119 inverse sqrt ( temp 2-component vector of float)
0:119 'inF0' ( in 2-component vector of float)
0:120 clamp ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:120 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:120 Constant:
0:120 0.000000
0:120 Constant:
0:120 1.000000
0:121 Sign ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:121 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:122 sine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:122 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:123 Sequence
0:123 move second child to first child ( temp 2-component vector of float)
0:123 'inF1' ( in 2-component vector of float)
0:123 sine ( temp 2-component vector of float)
0:123 'inF0' ( in 2-component vector of float)
0:123 move second child to first child ( temp 2-component vector of float)
0:123 'inF2' ( in 2-component vector of float)
0:123 cosine ( temp 2-component vector of float)
0:123 'inF0' ( in 2-component vector of float)
0:124 hyp. sine ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:124 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:125 smoothstep ( temp 2-component vector of float)
0:125 'inF0' ( in 2-component vector of float)
0:125 'inF1' ( in 2-component vector of float)
0:125 'inF2' ( in 2-component vector of float)
0:126 sqrt ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:126 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:127 step ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:127 'inF0' ( in 2-component vector of float)
0:127 'inF1' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:128 tangent ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:128 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:129 hyp. tangent ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:129 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:131 trunc ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:131 'inF0' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:134 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:134 Constant:
0:134 1.000000
0:134 2.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:138 Function Definition: VertexShaderFunction3(vf3;vf3;vf3;vu3;vu3; ( temp 3-component vector of float)
0:138 Function Parameters:
0:138 'inF0' ( in 3-component vector of float)
0:138 'inF1' ( in 3-component vector of float)
0:138 'inF2' ( in 3-component vector of float)
0:138 'inU0' ( in 3-component vector of uint)
0:138 'inU1' ( in 3-component vector of uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:139 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:139 Convert float to bool ( temp 3-component vector of bool)
0:139 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:140 Absolute value ( temp 3-component vector of float)
0:140 'inF0' ( in 3-component vector of float)
0:141 arc cosine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:141 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:142 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:142 Convert float to bool ( temp 3-component vector of bool)
0:142 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:143 arc sine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:143 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:144 floatBitsToInt ( temp 3-component vector of int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:144 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:145 floatBitsToUint ( temp 3-component vector of uint)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:145 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:146 intBitsToFloat ( temp 3-component vector of float)
0:146 'inU0' ( in 3-component vector of uint)
0:148 arc tangent ( temp 3-component vector of float)
0:148 'inF0' ( in 3-component vector of float)
0:149 arc tangent ( temp 3-component vector of float)
0:149 'inF0' ( in 3-component vector of float)
0:149 'inF1' ( in 3-component vector of float)
0:150 Ceiling ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:150 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:151 clamp ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:151 'inF0' ( in 3-component vector of float)
0:151 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:151 'inF2' ( in 3-component vector of float)
0:152 cosine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:152 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:153 hyp. cosine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:153 'inF0' ( in 3-component vector of float)
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:154 bitCount ( temp 3-component vector of int)
0:154 Constant:
0:154 7 (const int)
0:154 3 (const int)
0:154 5 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:155 cross-product ( temp 3-component vector of float)
0:155 'inF0' ( in 3-component vector of float)
0:155 'inF1' ( in 3-component vector of float)
0:156 degrees ( temp 3-component vector of float)
0:156 'inF0' ( in 3-component vector of float)
0:157 distance ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:157 'inF0' ( in 3-component vector of float)
0:157 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:158 dot-product ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:158 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:158 'inF1' ( in 3-component vector of float)
0:162 exp ( temp 3-component vector of float)
0:162 'inF0' ( in 3-component vector of float)
0:163 exp2 ( temp 3-component vector of float)
0:163 'inF0' ( in 3-component vector of float)
0:164 face-forward ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:164 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:164 'inF1' ( in 3-component vector of float)
0:164 'inF2' ( in 3-component vector of float)
0:165 findMSB ( temp int)
0:165 Constant:
0:165 7 (const int)
0:166 findLSB ( temp int)
0:166 Constant:
0:166 7 (const int)
0:167 Floor ( temp 3-component vector of float)
0:167 'inF0' ( in 3-component vector of float)
0:169 mod ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:169 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:169 'inF1' ( in 3-component vector of float)
0:170 Fraction ( temp 3-component vector of float)
0:170 'inF0' ( in 3-component vector of float)
0:171 isinf ( temp 3-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:171 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:172 isnan ( temp 3-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:172 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:173 ldexp ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:173 'inF0' ( in 3-component vector of float)
0:173 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:174 mix ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:174 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:174 'inF1' ( in 3-component vector of float)
0:174 'inF2' ( in 3-component vector of float)
0:175 length ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:175 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:176 log ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:176 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:177 vector-scale ( temp 3-component vector of float)
0:177 log2 ( temp 3-component vector of float)
0:177 'inF0' ( in 3-component vector of float)
0:177 Constant:
0:177 0.301030
0:178 log2 ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:178 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:179 max ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:179 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:179 'inF1' ( in 3-component vector of float)
0:180 min ( temp 3-component vector of float)
0:180 'inF0' ( in 3-component vector of float)
0:180 'inF1' ( in 3-component vector of float)
0:182 normalize ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:182 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:183 pow ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:183 'inF0' ( in 3-component vector of float)
0:183 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:184 radians ( temp 3-component vector of float)
0:184 'inF0' ( in 3-component vector of float)
0:185 reflect ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:185 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:185 'inF1' ( in 3-component vector of float)
0:186 refract ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:186 'inF0' ( in 3-component vector of float)
0:186 'inF1' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:186 Constant:
0:186 2.000000
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:187 bitFieldReverse ( temp 3-component vector of int)
0:187 Constant:
0:187 1 (const int)
0:187 2 (const int)
0:187 3 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:188 roundEven ( temp 3-component vector of float)
0:188 'inF0' ( in 3-component vector of float)
0:189 inverse sqrt ( temp 3-component vector of float)
0:189 'inF0' ( in 3-component vector of float)
0:190 clamp ( temp 3-component vector of float)
0:190 'inF0' ( in 3-component vector of float)
0:190 Constant:
0:190 0.000000
0:190 Constant:
0:190 1.000000
0:191 Sign ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:191 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:192 sine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:192 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:193 Sequence
0:193 move second child to first child ( temp 3-component vector of float)
0:193 'inF1' ( in 3-component vector of float)
0:193 sine ( temp 3-component vector of float)
0:193 'inF0' ( in 3-component vector of float)
0:193 move second child to first child ( temp 3-component vector of float)
0:193 'inF2' ( in 3-component vector of float)
0:193 cosine ( temp 3-component vector of float)
0:193 'inF0' ( in 3-component vector of float)
0:194 hyp. sine ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:194 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:195 smoothstep ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:195 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:195 'inF1' ( in 3-component vector of float)
0:195 'inF2' ( in 3-component vector of float)
0:196 sqrt ( temp 3-component vector of float)
0:196 'inF0' ( in 3-component vector of float)
0:197 step ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:197 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:197 'inF1' ( in 3-component vector of float)
0:198 tangent ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:198 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:199 hyp. tangent ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:199 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:201 trunc ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:201 'inF0' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:204 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:204 Constant:
0:204 1.000000
0:204 2.000000
0:204 3.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:208 Function Definition: VertexShaderFunction4(vf4;vf4;vf4;vu4;vu4; ( temp 4-component vector of float)
0:208 Function Parameters:
0:208 'inF0' ( in 4-component vector of float)
0:208 'inF1' ( in 4-component vector of float)
0:208 'inF2' ( in 4-component vector of float)
0:208 'inU0' ( in 4-component vector of uint)
0:208 'inU1' ( in 4-component vector of uint)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:209 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:209 Convert float to bool ( temp 4-component vector of bool)
0:209 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:210 Absolute value ( temp 4-component vector of float)
0:210 'inF0' ( in 4-component vector of float)
0:211 arc cosine ( temp 4-component vector of float)
0:211 'inF0' ( in 4-component vector of float)
0:212 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:212 Convert float to bool ( temp 4-component vector of bool)
0:212 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:213 arc sine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:213 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:214 floatBitsToInt ( temp 4-component vector of int)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:214 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:215 floatBitsToUint ( temp 4-component vector of uint)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:215 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:216 intBitsToFloat ( temp 4-component vector of float)
0:216 'inU0' ( in 4-component vector of uint)
0:218 arc tangent ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:218 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:219 arc tangent ( temp 4-component vector of float)
0:219 'inF0' ( in 4-component vector of float)
0:219 'inF1' ( in 4-component vector of float)
0:220 Ceiling ( temp 4-component vector of float)
0:220 'inF0' ( in 4-component vector of float)
0:221 clamp ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:221 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:221 'inF1' ( in 4-component vector of float)
0:221 'inF2' ( in 4-component vector of float)
0:222 cosine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:222 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:223 hyp. cosine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:223 'inF0' ( in 4-component vector of float)
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:224 bitCount ( temp 4-component vector of int)
0:224 Constant:
0:224 7 (const int)
0:224 3 (const int)
0:224 5 (const int)
0:224 2 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:225 degrees ( temp 4-component vector of float)
0:225 'inF0' ( in 4-component vector of float)
0:226 distance ( temp float)
0:226 'inF0' ( in 4-component vector of float)
0:226 'inF1' ( in 4-component vector of float)
0:227 dot-product ( temp float)
0:227 'inF0' ( in 4-component vector of float)
0:227 'inF1' ( in 4-component vector of float)
0:228 Construct vec4 ( temp 4-component vector of float)
0:228 Constant:
0:228 1.000000
0:228 component-wise multiply ( temp float)
0:228 direct index ( temp float)
0:228 'inF0' ( in 4-component vector of float)
0:228 Constant:
0:228 1 (const int)
0:228 direct index ( temp float)
0:228 'inF1' ( in 4-component vector of float)
0:228 Constant:
0:228 1 (const int)
0:228 direct index ( temp float)
0:228 'inF0' ( in 4-component vector of float)
0:228 Constant:
0:228 2 (const int)
0:228 direct index ( temp float)
0:228 'inF1' ( in 4-component vector of float)
0:228 Constant:
0:228 3 (const int)
0:232 exp ( temp 4-component vector of float)
0:232 'inF0' ( in 4-component vector of float)
0:233 exp2 ( temp 4-component vector of float)
0:233 'inF0' ( in 4-component vector of float)
0:234 face-forward ( temp 4-component vector of float)
0:234 'inF0' ( in 4-component vector of float)
0:234 'inF1' ( in 4-component vector of float)
0:234 'inF2' ( in 4-component vector of float)
0:235 findMSB ( temp int)
0:235 Constant:
0:235 7 (const int)
0:236 findLSB ( temp int)
0:236 Constant:
0:236 7 (const int)
0:237 Floor ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:237 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:239 mod ( temp 4-component vector of float)
0:239 'inF0' ( in 4-component vector of float)
0:239 'inF1' ( in 4-component vector of float)
0:240 Fraction ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:240 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:241 isinf ( temp 4-component vector of bool)
0:241 'inF0' ( in 4-component vector of float)
0:242 isnan ( temp 4-component vector of bool)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:242 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:243 ldexp ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:243 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:243 'inF1' ( in 4-component vector of float)
0:244 mix ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:244 'inF0' ( in 4-component vector of float)
0:244 'inF1' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:244 'inF2' ( in 4-component vector of float)
0:245 length ( temp float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:245 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:246 log ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:246 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:247 vector-scale ( temp 4-component vector of float)
0:247 log2 ( temp 4-component vector of float)
0:247 'inF0' ( in 4-component vector of float)
0:247 Constant:
0:247 0.301030
0:248 log2 ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:248 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:249 max ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:249 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:249 'inF1' ( in 4-component vector of float)
0:250 min ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:250 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:250 'inF1' ( in 4-component vector of float)
0:252 normalize ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:252 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:253 pow ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:253 'inF0' ( in 4-component vector of float)
0:253 'inF1' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:254 radians ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:254 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:255 reflect ( temp 4-component vector of float)
0:255 'inF0' ( in 4-component vector of float)
0:255 'inF1' ( in 4-component vector of float)
0:256 refract ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:256 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:256 'inF1' ( in 4-component vector of float)
0:256 Constant:
0:256 2.000000
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:257 bitFieldReverse ( temp 4-component vector of int)
0:257 Constant:
0:257 1 (const int)
0:257 2 (const int)
0:257 3 (const int)
0:257 4 (const int)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:258 roundEven ( temp 4-component vector of float)
0:258 'inF0' ( in 4-component vector of float)
0:259 inverse sqrt ( temp 4-component vector of float)
0:259 'inF0' ( in 4-component vector of float)
0:260 clamp ( temp 4-component vector of float)
0:260 'inF0' ( in 4-component vector of float)
0:260 Constant:
0:260 0.000000
0:260 Constant:
0:260 1.000000
0:261 Sign ( temp 4-component vector of float)
0:261 'inF0' ( in 4-component vector of float)
0:262 sine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:262 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:263 Sequence
0:263 move second child to first child ( temp 4-component vector of float)
0:263 'inF1' ( in 4-component vector of float)
0:263 sine ( temp 4-component vector of float)
0:263 'inF0' ( in 4-component vector of float)
0:263 move second child to first child ( temp 4-component vector of float)
0:263 'inF2' ( in 4-component vector of float)
0:263 cosine ( temp 4-component vector of float)
0:263 'inF0' ( in 4-component vector of float)
0:264 hyp. sine ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:264 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:265 smoothstep ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:265 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:265 'inF1' ( in 4-component vector of float)
0:265 'inF2' ( in 4-component vector of float)
0:266 sqrt ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:266 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:267 step ( temp 4-component vector of float)
0:267 'inF0' ( in 4-component vector of float)
0:267 'inF1' ( in 4-component vector of float)
0:268 tangent ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:268 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:269 hyp. tangent ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:269 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:271 trunc ( temp 4-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:271 'inF0' ( in 4-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:274 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:274 Constant:
0:274 1.000000
0:274 2.000000
0:274 3.000000
0:274 4.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:331 Function Definition: VertexShaderFunction2x2(mf22;mf22;mf22; ( temp 2X2 matrix of float)
0:331 Function Parameters:
0:331 'inF0' ( in 2X2 matrix of float)
0:331 'inF1' ( in 2X2 matrix of float)
0:331 'inF2' ( in 2X2 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:333 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:333 Convert float to bool ( temp 2X2 matrix of bool)
0:333 'inF0' ( in 2X2 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:333 Absolute value ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 arc cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:333 Convert float to bool ( temp 2X2 matrix of bool)
0:333 'inF0' ( in 2X2 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:333 arc sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 arc tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 arc tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 Ceiling ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 clamp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 hyp. cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 degrees ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 determinant ( temp float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 exp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 exp2 ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 findMSB ( temp int)
0:333 Constant:
0:333 7 (const int)
0:333 findLSB ( temp int)
0:333 Constant:
0:333 7 (const int)
0:333 Floor ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 mod ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 Fraction ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 ldexp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 mix ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 log ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 matrix-scale ( temp 2X2 matrix of float)
0:333 log2 ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 Constant:
0:333 0.301030
0:333 log2 ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 max ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 min ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 pow ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 radians ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 roundEven ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 inverse sqrt ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 clamp ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 Constant:
0:333 0.000000
0:333 Constant:
0:333 1.000000
0:333 Sign ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 Sequence
0:333 move second child to first child ( temp 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 move second child to first child ( temp 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 cosine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 hyp. sine ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 smoothstep ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 'inF2' ( in 2X2 matrix of float)
0:333 sqrt ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 step ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 'inF1' ( in 2X2 matrix of float)
0:333 tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 hyp. tangent ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 transpose ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:333 trunc ( temp 2X2 matrix of float)
0:333 'inF0' ( in 2X2 matrix of float)
0:336 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:336 Constant:
0:336 2.000000
0:336 2.000000
0:336 2.000000
0:336 2.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:340 Function Definition: VertexShaderFunction3x3(mf33;mf33;mf33; ( temp 3X3 matrix of float)
0:340 Function Parameters:
0:340 'inF0' ( in 3X3 matrix of float)
0:340 'inF1' ( in 3X3 matrix of float)
0:340 'inF2' ( in 3X3 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:342 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:342 Convert float to bool ( temp 3X3 matrix of bool)
0:342 'inF0' ( in 3X3 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:342 Absolute value ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 arc cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:342 Convert float to bool ( temp 3X3 matrix of bool)
0:342 'inF0' ( in 3X3 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:342 arc sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 arc tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 arc tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 Ceiling ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 clamp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 hyp. cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 degrees ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 determinant ( temp float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 exp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 exp2 ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 findMSB ( temp int)
0:342 Constant:
0:342 7 (const int)
0:342 findLSB ( temp int)
0:342 Constant:
0:342 7 (const int)
0:342 Floor ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 mod ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 Fraction ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 ldexp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 mix ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 log ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 matrix-scale ( temp 3X3 matrix of float)
0:342 log2 ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 Constant:
0:342 0.301030
0:342 log2 ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 max ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 min ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 pow ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 radians ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 roundEven ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 inverse sqrt ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 clamp ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 Constant:
0:342 0.000000
0:342 Constant:
0:342 1.000000
0:342 Sign ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 Sequence
0:342 move second child to first child ( temp 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 move second child to first child ( temp 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 cosine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 hyp. sine ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 smoothstep ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 'inF2' ( in 3X3 matrix of float)
0:342 sqrt ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 step ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 'inF1' ( in 3X3 matrix of float)
0:342 tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 hyp. tangent ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 transpose ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:342 trunc ( temp 3X3 matrix of float)
0:342 'inF0' ( in 3X3 matrix of float)
0:345 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:345 Constant:
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
0:345 3.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:349 Function Definition: VertexShaderFunction4x4(mf44;mf44;mf44; ( temp 4X4 matrix of float)
0:349 Function Parameters:
0:349 'inF0' ( in 4X4 matrix of float)
0:349 'inF1' ( in 4X4 matrix of float)
0:349 'inF2' ( in 4X4 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:351 all ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:351 Convert float to bool ( temp 4X4 matrix of bool)
0:351 'inF0' ( in 4X4 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:351 Absolute value ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 arc cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 any ( temp bool)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
0:351 Convert float to bool ( temp 4X4 matrix of bool)
0:351 'inF0' ( in 4X4 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:351 arc sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 arc tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 arc tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 Ceiling ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 clamp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 hyp. cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 degrees ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 determinant ( temp float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 exp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 exp2 ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 findMSB ( temp int)
0:351 Constant:
0:351 7 (const int)
0:351 findLSB ( temp int)
0:351 Constant:
0:351 7 (const int)
0:351 Floor ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 mod ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 Fraction ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 ldexp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 mix ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 log ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 matrix-scale ( temp 4X4 matrix of float)
0:351 log2 ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 Constant:
0:351 0.301030
0:351 log2 ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 max ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 min ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 pow ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 radians ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 roundEven ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 inverse sqrt ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 clamp ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 Constant:
0:351 0.000000
0:351 Constant:
0:351 1.000000
0:351 Sign ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 Sequence
0:351 move second child to first child ( temp 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 move second child to first child ( temp 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 cosine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 hyp. sine ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 smoothstep ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 'inF2' ( in 4X4 matrix of float)
0:351 sqrt ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 step ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 'inF1' ( in 4X4 matrix of float)
0:351 tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 hyp. tangent ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 transpose ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:351 trunc ( temp 4X4 matrix of float)
0:351 'inF0' ( in 4X4 matrix of float)
0:354 Branch: Return with expression
fix error message for hlslGrammar::acceptConstructor Fix the error message for when an erroneous HLSL constructor statement is detected. Prior to this change, such error messages would not show correct file path and line number information. Additionally, update test data to account for updated error messages.
2021-05-11 08:42:11 +00:00
0:354 Constant:
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
0:354 4.000000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:372 Function Definition: TestGenMul2(f1;f1;vf2;vf2;mf22;mf22; ( temp void)
0:372 Function Parameters:
0:372 'inF0' ( in float)
0:372 'inF1' ( in float)
0:372 'inFV0' ( in 2-component vector of float)
0:372 'inFV1' ( in 2-component vector of float)
0:372 'inFM0' ( in 2X2 matrix of float)
0:372 'inFM1' ( in 2X2 matrix of float)
Add lerp, fix sincos return type, ret type tests, non-square mats, HLSL->AST tx semantics
2016-06-20 19:26:59 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:373 Sequence
0:373 move second child to first child ( temp float)
0:373 'r0' ( temp float)
0:373 component-wise multiply ( temp float)
0:373 'inF1' ( in float)
0:373 'inF0' ( in float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r1' ( temp 2-component vector of float)
0:373 vector-scale ( temp 2-component vector of float)
0:373 'inF0' ( in float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r2' ( temp 2-component vector of float)
0:373 vector-scale ( temp 2-component vector of float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 'inF0' ( in float)
0:373 Sequence
0:373 move second child to first child ( temp float)
0:373 'r3' ( temp float)
0:373 dot-product ( temp float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 'inFV1' ( in 2-component vector of float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r4' ( temp 2-component vector of float)
0:373 vector-times-matrix ( temp 2-component vector of float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 Sequence
0:373 move second child to first child ( temp 2-component vector of float)
0:373 'r5' ( temp 2-component vector of float)
0:373 matrix-times-vector ( temp 2-component vector of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 'inFV0' ( in 2-component vector of float)
0:373 Sequence
0:373 move second child to first child ( temp 2X2 matrix of float)
0:373 'r6' ( temp 2X2 matrix of float)
0:373 matrix-scale ( temp 2X2 matrix of float)
0:373 'inF0' ( in float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 Sequence
0:373 move second child to first child ( temp 2X2 matrix of float)
0:373 'r7' ( temp 2X2 matrix of float)
0:373 matrix-scale ( temp 2X2 matrix of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:373 'inF0' ( in float)
0:373 Sequence
0:373 move second child to first child ( temp 2X2 matrix of float)
0:373 'r8' ( temp 2X2 matrix of float)
0:373 matrix-multiply ( temp 2X2 matrix of float)
0:373 'inFM1' ( in 2X2 matrix of float)
0:373 'inFM0' ( in 2X2 matrix of float)
0:379 Function Definition: TestGenMul3(f1;f1;vf3;vf3;mf33;mf33; ( temp void)
0:379 Function Parameters:
0:379 'inF0' ( in float)
0:379 'inF1' ( in float)
0:379 'inFV0' ( in 3-component vector of float)
0:379 'inFV1' ( in 3-component vector of float)
0:379 'inFM0' ( in 3X3 matrix of float)
0:379 'inFM1' ( in 3X3 matrix of float)
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:380 Sequence
0:380 move second child to first child ( temp float)
0:380 'r0' ( temp float)
0:380 component-wise multiply ( temp float)
0:380 'inF1' ( in float)
0:380 'inF0' ( in float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r1' ( temp 3-component vector of float)
0:380 vector-scale ( temp 3-component vector of float)
0:380 'inF0' ( in float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r2' ( temp 3-component vector of float)
0:380 vector-scale ( temp 3-component vector of float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 'inF0' ( in float)
0:380 Sequence
0:380 move second child to first child ( temp float)
0:380 'r3' ( temp float)
0:380 dot-product ( temp float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 'inFV1' ( in 3-component vector of float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r4' ( temp 3-component vector of float)
0:380 vector-times-matrix ( temp 3-component vector of float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 Sequence
0:380 move second child to first child ( temp 3-component vector of float)
0:380 'r5' ( temp 3-component vector of float)
0:380 matrix-times-vector ( temp 3-component vector of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 'inFV0' ( in 3-component vector of float)
0:380 Sequence
0:380 move second child to first child ( temp 3X3 matrix of float)
0:380 'r6' ( temp 3X3 matrix of float)
0:380 matrix-scale ( temp 3X3 matrix of float)
0:380 'inF0' ( in float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 Sequence
0:380 move second child to first child ( temp 3X3 matrix of float)
0:380 'r7' ( temp 3X3 matrix of float)
0:380 matrix-scale ( temp 3X3 matrix of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:380 'inF0' ( in float)
0:380 Sequence
0:380 move second child to first child ( temp 3X3 matrix of float)
0:380 'r8' ( temp 3X3 matrix of float)
0:380 matrix-multiply ( temp 3X3 matrix of float)
0:380 'inFM1' ( in 3X3 matrix of float)
0:380 'inFM0' ( in 3X3 matrix of float)
0:386 Function Definition: TestGenMul4(f1;f1;vf4;vf4;mf44;mf44; ( temp void)
0:386 Function Parameters:
0:386 'inF0' ( in float)
0:386 'inF1' ( in float)
0:386 'inFV0' ( in 4-component vector of float)
0:386 'inFV1' ( in 4-component vector of float)
0:386 'inFM0' ( in 4X4 matrix of float)
0:386 'inFM1' ( in 4X4 matrix of float)
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:387 Sequence
0:387 move second child to first child ( temp float)
0:387 'r0' ( temp float)
0:387 component-wise multiply ( temp float)
0:387 'inF1' ( in float)
0:387 'inF0' ( in float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r1' ( temp 4-component vector of float)
0:387 vector-scale ( temp 4-component vector of float)
0:387 'inF0' ( in float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r2' ( temp 4-component vector of float)
0:387 vector-scale ( temp 4-component vector of float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 'inF0' ( in float)
0:387 Sequence
0:387 move second child to first child ( temp float)
0:387 'r3' ( temp float)
0:387 dot-product ( temp float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 'inFV1' ( in 4-component vector of float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r4' ( temp 4-component vector of float)
0:387 vector-times-matrix ( temp 4-component vector of float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 Sequence
0:387 move second child to first child ( temp 4-component vector of float)
0:387 'r5' ( temp 4-component vector of float)
0:387 matrix-times-vector ( temp 4-component vector of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 'inFV0' ( in 4-component vector of float)
0:387 Sequence
0:387 move second child to first child ( temp 4X4 matrix of float)
0:387 'r6' ( temp 4X4 matrix of float)
0:387 matrix-scale ( temp 4X4 matrix of float)
0:387 'inF0' ( in float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 Sequence
0:387 move second child to first child ( temp 4X4 matrix of float)
0:387 'r7' ( temp 4X4 matrix of float)
0:387 matrix-scale ( temp 4X4 matrix of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:387 'inF0' ( in float)
0:387 Sequence
0:387 move second child to first child ( temp 4X4 matrix of float)
0:387 'r8' ( temp 4X4 matrix of float)
0:387 matrix-multiply ( temp 4X4 matrix of float)
0:387 'inFM1' ( in 4X4 matrix of float)
0:387 'inFM0' ( in 4X4 matrix of float)
0:396 Function Definition: TestGenMulNxM(f1;f1;vf2;vf3;mf23;mf32;mf33;mf34;mf24; ( temp void)
0:396 Function Parameters:
0:396 'inF0' ( in float)
0:396 'inF1' ( in float)
0:396 'inFV2' ( in 2-component vector of float)
0:396 'inFV3' ( in 3-component vector of float)
0:396 'inFM2x3' ( in 2X3 matrix of float)
0:396 'inFM3x2' ( in 3X2 matrix of float)
0:396 'inFM3x3' ( in 3X3 matrix of float)
0:396 'inFM3x4' ( in 3X4 matrix of float)
0:396 'inFM2x4' ( in 2X4 matrix of float)
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
0:? Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:397 Sequence
0:397 move second child to first child ( temp float)
0:397 'r00' ( temp float)
0:397 component-wise multiply ( temp float)
0:397 'inF1' ( in float)
0:397 'inF0' ( in float)
0:398 Sequence
0:398 move second child to first child ( temp 2-component vector of float)
0:398 'r01' ( temp 2-component vector of float)
0:398 vector-scale ( temp 2-component vector of float)
0:398 'inF0' ( in float)
0:398 'inFV2' ( in 2-component vector of float)
0:399 Sequence
0:399 move second child to first child ( temp 3-component vector of float)
0:399 'r02' ( temp 3-component vector of float)
0:399 vector-scale ( temp 3-component vector of float)
0:399 'inF0' ( in float)
0:399 'inFV3' ( in 3-component vector of float)
0:400 Sequence
0:400 move second child to first child ( temp 2-component vector of float)
0:400 'r03' ( temp 2-component vector of float)
0:400 vector-scale ( temp 2-component vector of float)
0:400 'inFV2' ( in 2-component vector of float)
0:400 'inF0' ( in float)
0:401 Sequence
0:401 move second child to first child ( temp 3-component vector of float)
0:401 'r04' ( temp 3-component vector of float)
0:401 vector-scale ( temp 3-component vector of float)
0:401 'inFV3' ( in 3-component vector of float)
0:401 'inF0' ( in float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:402 Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:402 move second child to first child ( temp float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:402 'r05' ( temp float)
0:402 dot-product ( temp float)
0:402 'inFV2' ( in 2-component vector of float)
0:402 'inFV2' ( in 2-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:403 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:403 move second child to first child ( temp float)
0:403 'r06' ( temp float)
0:403 dot-product ( temp float)
0:403 'inFV3' ( in 3-component vector of float)
0:403 'inFV3' ( in 3-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:404 Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:404 move second child to first child ( temp 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:404 'r07' ( temp 3-component vector of float)
0:404 matrix-times-vector ( temp 3-component vector of float)
0:404 'inFM2x3' ( in 2X3 matrix of float)
0:404 'inFV2' ( in 2-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:405 Sequence
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:405 move second child to first child ( temp 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:405 'r08' ( temp 2-component vector of float)
0:405 matrix-times-vector ( temp 2-component vector of float)
0:405 'inFM3x2' ( in 3X2 matrix of float)
0:405 'inFV3' ( in 3-component vector of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:406 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:406 move second child to first child ( temp 2-component vector of float)
0:406 'r09' ( temp 2-component vector of float)
0:406 vector-times-matrix ( temp 2-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:406 'inFV3' ( in 3-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:406 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:407 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:407 move second child to first child ( temp 3-component vector of float)
0:407 'r10' ( temp 3-component vector of float)
0:407 vector-times-matrix ( temp 3-component vector of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:407 'inFV2' ( in 2-component vector of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:407 'inFM3x2' ( in 3X2 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:408 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:408 move second child to first child ( temp 2X3 matrix of float)
0:408 'r11' ( temp 2X3 matrix of float)
0:408 matrix-scale ( temp 2X3 matrix of float)
0:408 'inF0' ( in float)
0:408 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:409 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:409 move second child to first child ( temp 3X2 matrix of float)
0:409 'r12' ( temp 3X2 matrix of float)
0:409 matrix-scale ( temp 3X2 matrix of float)
0:409 'inF0' ( in float)
0:409 'inFM3x2' ( in 3X2 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:410 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:410 move second child to first child ( temp 2X2 matrix of float)
0:410 'r13' ( temp 2X2 matrix of float)
0:410 matrix-multiply ( temp 2X2 matrix of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:410 'inFM3x2' ( in 3X2 matrix of float)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:410 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:411 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:411 move second child to first child ( temp 2X3 matrix of float)
0:411 'r14' ( temp 2X3 matrix of float)
0:411 matrix-multiply ( temp 2X3 matrix of float)
0:411 'inFM3x3' ( in 3X3 matrix of float)
GLSL output: Removed fixed-size buffer; fixes #769. Makes some white-space differences in most output, plus a few cases where more could have been put out but was cut short by the previous fix-sized buffer.
2017-03-15 01:43:13 +00:00
0:411 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:412 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:412 move second child to first child ( temp 2X4 matrix of float)
0:412 'r15' ( temp 2X4 matrix of float)
0:412 matrix-multiply ( temp 2X4 matrix of float)
0:412 'inFM3x4' ( in 3X4 matrix of float)
0:412 'inFM2x3' ( in 2X3 matrix of float)
HLSL: Grammar: Make comma-separated declaration lists work.
2016-07-02 01:06:44 +00:00
0:413 Sequence
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
0:413 move second child to first child ( temp 3X4 matrix of float)
0:413 'r16' ( temp 3X4 matrix of float)
0:413 matrix-multiply ( temp 3X4 matrix of float)
0:413 'inFM2x4' ( in 2X4 matrix of float)
0:413 'inFM3x2' ( in 3X2 matrix of float)
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
0:? Linker Objects
Modify testing to only record validation pass/fail * Changed unit tests to only record known the validation pass/fail status * errors are output as part of the failure message if the result is unexpected * can turn off validation for each test individually * Moved some SPV_KHR_vulkan_memory_model tests to be compiled for Vulkan 1.1
2018-12-04 15:43:23 +00:00
Validation failed
SPV Version: Emit the requested SPIR-V version, not the header version. Fixes #1236.
2018-02-01 01:35:56 +00:00
// Module Version 10000
Update CHANGES for release 11.12.0
2022-10-13 00:16:41 +00:00
// Generated by (magic number): 8000b
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
// Id's are bound by 1225
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
Capability Shader
1: ExtInstImport "GLSL.std.450"
MemoryModel Logical GLSL450
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
EntryPoint Vertex 4 "VertexShaderFunction"
HLSL: Emit the OpSource HLSL instruction for HLSL, using new headers.
2017-04-07 21:33:08 +00:00
Source HLSL 500
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
Name 4 "VertexShaderFunction"
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
Name 16 "VertexShaderFunctionS(f1;f1;f1;u1;u1;"
Name 11 "inF0"
Name 12 "inF1"
Name 13 "inF2"
Name 14 "inU0"
Name 15 "inU1"
Name 22 "VertexShaderFunction1(vf1;vf1;vf1;"
Name 19 "inF0"
Name 20 "inF1"
Name 21 "inF2"
Name 34 "VertexShaderFunction2(vf2;vf2;vf2;vu2;vu2;"
Name 29 "inF0"
Name 30 "inF1"
Name 31 "inF2"
Name 32 "inU0"
Name 33 "inU1"
Name 46 "VertexShaderFunction3(vf3;vf3;vf3;vu3;vu3;"
Name 41 "inF0"
Name 42 "inF1"
Name 43 "inF2"
Name 44 "inU0"
Name 45 "inU1"
Name 58 "VertexShaderFunction4(vf4;vf4;vf4;vu4;vu4;"
Name 53 "inF0"
Name 54 "inF1"
Name 55 "inF2"
Name 56 "inU0"
Name 57 "inU1"
Name 66 "VertexShaderFunction2x2(mf22;mf22;mf22;"
Name 63 "inF0"
Name 64 "inF1"
Name 65 "inF2"
Name 74 "VertexShaderFunction3x3(mf33;mf33;mf33;"
Name 71 "inF0"
Name 72 "inF1"
Name 73 "inF2"
Name 82 "VertexShaderFunction4x4(mf44;mf44;mf44;"
Name 79 "inF0"
Name 80 "inF1"
Name 81 "inF2"
Name 91 "TestGenMul2(f1;f1;vf2;vf2;mf22;mf22;"
Name 85 "inF0"
Name 86 "inF1"
Name 87 "inFV0"
Name 88 "inFV1"
Name 89 "inFM0"
Name 90 "inFM1"
Name 100 "TestGenMul3(f1;f1;vf3;vf3;mf33;mf33;"
Name 94 "inF0"
Name 95 "inF1"
Name 96 "inFV0"
Name 97 "inFV1"
Name 98 "inFM0"
Name 99 "inFM1"
Name 109 "TestGenMul4(f1;f1;vf4;vf4;mf44;mf44;"
Name 103 "inF0"
Name 104 "inF1"
Name 105 "inFV0"
Name 106 "inFV1"
Name 107 "inFM0"
Name 108 "inFM1"
WIP: HLSL: Treat HLSL rows as GLSL columns. WIP: HLSL: EOpGenMul arg reversal
2016-08-24 20:36:13 +00:00
Name 129 "TestGenMulNxM(f1;f1;vf2;vf3;mf23;mf32;mf33;mf34;mf24;"
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
Name 120 "inF0"
Name 121 "inF1"
Name 122 "inFV2"
Name 123 "inFV3"
Name 124 "inFM2x3"
Name 125 "inFM3x2"
Name 126 "inFM3x3"
Name 127 "inFM3x4"
Name 128 "inFM2x4"
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
Name 1049 "r0"
Name 1053 "r1"
Name 1057 "r2"
Name 1061 "r3"
Name 1065 "r4"
Name 1069 "r5"
Name 1073 "r6"
Name 1077 "r7"
Name 1081 "r8"
Name 1085 "r0"
Name 1089 "r1"
Name 1093 "r2"
Name 1097 "r3"
Name 1101 "r4"
Name 1105 "r5"
Name 1109 "r6"
Name 1113 "r7"
Name 1117 "r8"
Name 1121 "r0"
Name 1125 "r1"
Name 1129 "r2"
Name 1133 "r3"
Name 1137 "r4"
Name 1141 "r5"
Name 1145 "r6"
Name 1149 "r7"
Name 1153 "r8"
Name 1157 "r00"
Name 1161 "r01"
Name 1165 "r02"
Name 1169 "r03"
Name 1173 "r04"
Name 1177 "r05"
Name 1181 "r06"
Name 1185 "r07"
Name 1189 "r08"
Name 1193 "r09"
Name 1197 "r10"
Name 1201 "r11"
Name 1205 "r12"
Name 1209 "r13"
Name 1213 "r14"
Name 1217 "r15"
Name 1221 "r16"
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
2: TypeVoid
3: TypeFunction 2
6: TypeFloat 32
7: TypePointer Function 6(float)
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
8: TypeInt 32 0
9: TypePointer Function 8(int)
10: TypeFunction 6(float) 7(ptr) 7(ptr) 7(ptr) 9(ptr) 9(ptr)
18: TypeFunction 6(float) 7(ptr) 7(ptr) 7(ptr)
24: TypeVector 6(float) 2
25: TypePointer Function 24(fvec2)
26: TypeVector 8(int) 2
27: TypePointer Function 26(ivec2)
28: TypeFunction 24(fvec2) 25(ptr) 25(ptr) 25(ptr) 27(ptr) 27(ptr)
36: TypeVector 6(float) 3
37: TypePointer Function 36(fvec3)
38: TypeVector 8(int) 3
39: TypePointer Function 38(ivec3)
40: TypeFunction 36(fvec3) 37(ptr) 37(ptr) 37(ptr) 39(ptr) 39(ptr)
48: TypeVector 6(float) 4
49: TypePointer Function 48(fvec4)
50: TypeVector 8(int) 4
51: TypePointer Function 50(ivec4)
52: TypeFunction 48(fvec4) 49(ptr) 49(ptr) 49(ptr) 51(ptr) 51(ptr)
60: TypeMatrix 24(fvec2) 2
61: TypePointer Function 60
62: TypeFunction 60 61(ptr) 61(ptr) 61(ptr)
68: TypeMatrix 36(fvec3) 3
69: TypePointer Function 68
70: TypeFunction 68 69(ptr) 69(ptr) 69(ptr)
76: TypeMatrix 48(fvec4) 4
77: TypePointer Function 76
78: TypeFunction 76 77(ptr) 77(ptr) 77(ptr)
84: TypeFunction 2 7(ptr) 7(ptr) 25(ptr) 25(ptr) 61(ptr) 61(ptr)
93: TypeFunction 2 7(ptr) 7(ptr) 37(ptr) 37(ptr) 69(ptr) 69(ptr)
102: TypeFunction 2 7(ptr) 7(ptr) 49(ptr) 49(ptr) 77(ptr) 77(ptr)
WIP: HLSL: Treat HLSL rows as GLSL columns. WIP: HLSL: EOpGenMul arg reversal
2016-08-24 20:36:13 +00:00
111: TypeMatrix 36(fvec3) 2
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
112: TypePointer Function 111
WIP: HLSL: Treat HLSL rows as GLSL columns. WIP: HLSL: EOpGenMul arg reversal
2016-08-24 20:36:13 +00:00
113: TypeMatrix 24(fvec2) 3
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
114: TypePointer Function 113
WIP: HLSL: Treat HLSL rows as GLSL columns. WIP: HLSL: EOpGenMul arg reversal
2016-08-24 20:36:13 +00:00
115: TypeMatrix 48(fvec4) 3
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
116: TypePointer Function 115
WIP: HLSL: Treat HLSL rows as GLSL columns. WIP: HLSL: EOpGenMul arg reversal
2016-08-24 20:36:13 +00:00
117: TypeMatrix 48(fvec4) 2
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
118: TypePointer Function 117
119: TypeFunction 2 7(ptr) 7(ptr) 25(ptr) 37(ptr) 112(ptr) 114(ptr) 69(ptr) 116(ptr) 118(ptr)
132: TypeBool
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
133: 6(float) Constant 0
146: TypeInt 32 1
167: 146(int) Constant 7
199: 6(float) Constant 1050288283
214: 146(int) Constant 2
221: 6(float) Constant 1065353216
253: TypeVector 132(bool) 2
254: 24(fvec2) ConstantComposite 133 133
267: TypeVector 146(int) 2
288: 146(int) Constant 3
289: 267(ivec2) ConstantComposite 167 288
354: 6(float) Constant 1073741824
356: 146(int) Constant 1
357: 267(ivec2) ConstantComposite 356 214
392: 24(fvec2) ConstantComposite 221 354
396: TypeVector 132(bool) 3
397: 36(fvec3) ConstantComposite 133 133 133
410: TypeVector 146(int) 3
431: 146(int) Constant 5
432: 410(ivec3) ConstantComposite 167 288 431
501: 410(ivec3) ConstantComposite 356 214 288
536: 6(float) Constant 1077936128
537: 36(fvec3) ConstantComposite 221 354 536
541: TypeVector 132(bool) 4
542: 48(fvec4) ConstantComposite 133 133 133 133
555: TypeVector 146(int) 4
576: 555(ivec4) ConstantComposite 167 288 431 214
586: 8(int) Constant 1
592: 8(int) Constant 2
595: 8(int) Constant 3
655: 146(int) Constant 4
656: 555(ivec4) ConstantComposite 356 214 288 655
691: 6(float) Constant 1082130432
692: 48(fvec4) ConstantComposite 221 354 536 691
696: TypeMatrix 253(bvec2) 2
806: 24(fvec2) ConstantComposite 354 354
807: 60 ConstantComposite 806 806
811: TypeMatrix 396(bvec3) 3
924: 36(fvec3) ConstantComposite 536 536 536
925: 68 ConstantComposite 924 924 924
929: TypeMatrix 541(bvec4) 4
1045: 48(fvec4) ConstantComposite 691 691 691 691
1046: 76 ConstantComposite 1045 1045 1045 1045
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
4(VertexShaderFunction): 2 Function None 3
5: Label
Always correctly terminate main. Issue #588, PR #600.
2016-11-26 20:23:20 +00:00
Return
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
FunctionEnd
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
16(VertexShaderFunctionS(f1;f1;f1;u1;u1;): 6(float) Function None 10
11(inF0): 7(ptr) FunctionParameter
12(inF1): 7(ptr) FunctionParameter
13(inF2): 7(ptr) FunctionParameter
14(inU0): 9(ptr) FunctionParameter
15(inU1): 9(ptr) FunctionParameter
17: Label
131: 6(float) Load 11(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
134: 132(bool) FUnordNotEqual 131 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
135: 132(bool) All 134
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
136: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
137: 6(float) ExtInst 1(GLSL.std.450) 4(FAbs) 136
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
138: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
139: 6(float) ExtInst 1(GLSL.std.450) 17(Acos) 138
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
140: 6(float) Load 11(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
141: 132(bool) FUnordNotEqual 140 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
142: 132(bool) Any 141
143: 6(float) Load 11(inF0)
144: 6(float) ExtInst 1(GLSL.std.450) 16(Asin) 143
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
145: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
147: 146(int) Bitcast 145
148: 6(float) Load 11(inF0)
149: 8(int) Bitcast 148
150: 8(int) Load 14(inU0)
151: 6(float) Bitcast 150
152: 6(float) Load 11(inF0)
153: 6(float) ExtInst 1(GLSL.std.450) 18(Atan) 152
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
154: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
155: 6(float) Load 12(inF1)
156: 6(float) ExtInst 1(GLSL.std.450) 25(Atan2) 154 155
157: 6(float) Load 11(inF0)
158: 6(float) ExtInst 1(GLSL.std.450) 9(Ceil) 157
159: 6(float) Load 11(inF0)
160: 6(float) Load 12(inF1)
161: 6(float) Load 13(inF2)
162: 6(float) ExtInst 1(GLSL.std.450) 43(FClamp) 159 160 161
163: 6(float) Load 11(inF0)
164: 6(float) ExtInst 1(GLSL.std.450) 14(Cos) 163
165: 6(float) Load 11(inF0)
166: 6(float) ExtInst 1(GLSL.std.450) 20(Cosh) 165
168: 146(int) BitCount 167
169: 6(float) Load 11(inF0)
170: 6(float) ExtInst 1(GLSL.std.450) 12(Degrees) 169
171: 6(float) Load 11(inF0)
172: 6(float) ExtInst 1(GLSL.std.450) 27(Exp) 171
173: 6(float) Load 11(inF0)
174: 6(float) ExtInst 1(GLSL.std.450) 29(Exp2) 173
175: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
176: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
177: 6(float) Load 11(inF0)
178: 6(float) ExtInst 1(GLSL.std.450) 8(Floor) 177
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
179: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
180: 6(float) Load 12(inF1)
181: 6(float) FMod 179 180
182: 6(float) Load 11(inF0)
183: 6(float) ExtInst 1(GLSL.std.450) 10(Fract) 182
184: 6(float) Load 11(inF0)
185: 132(bool) IsInf 184
186: 6(float) Load 11(inF0)
187: 132(bool) IsNan 186
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
188: 6(float) Load 11(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
189: 6(float) Load 12(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
190: 6(float) ExtInst 1(GLSL.std.450) 53(Ldexp) 188 189
191: 6(float) Load 11(inF0)
192: 6(float) Load 12(inF1)
193: 6(float) Load 13(inF2)
194: 6(float) ExtInst 1(GLSL.std.450) 46(FMix) 191 192 193
195: 6(float) Load 11(inF0)
196: 6(float) ExtInst 1(GLSL.std.450) 28(Log) 195
197: 6(float) Load 11(inF0)
198: 6(float) ExtInst 1(GLSL.std.450) 30(Log2) 197
200: 6(float) FMul 198 199
201: 6(float) Load 11(inF0)
202: 6(float) ExtInst 1(GLSL.std.450) 30(Log2) 201
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
203: 6(float) Load 11(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
204: 6(float) Load 12(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
205: 6(float) ExtInst 1(GLSL.std.450) 40(FMax) 203 204
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
206: 6(float) Load 11(inF0)
207: 6(float) Load 12(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
208: 6(float) ExtInst 1(GLSL.std.450) 37(FMin) 206 207
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
209: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
210: 6(float) Load 12(inF1)
211: 6(float) ExtInst 1(GLSL.std.450) 26(Pow) 209 210
212: 6(float) Load 11(inF0)
213: 6(float) ExtInst 1(GLSL.std.450) 11(Radians) 212
215: 146(int) BitReverse 214
216: 6(float) Load 11(inF0)
217: 6(float) ExtInst 1(GLSL.std.450) 2(RoundEven) 216
218: 6(float) Load 11(inF0)
219: 6(float) ExtInst 1(GLSL.std.450) 32(InverseSqrt) 218
220: 6(float) Load 11(inF0)
222: 6(float) ExtInst 1(GLSL.std.450) 43(FClamp) 220 133 221
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
223: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
224: 6(float) ExtInst 1(GLSL.std.450) 6(FSign) 223
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
225: 6(float) Load 11(inF0)
226: 6(float) ExtInst 1(GLSL.std.450) 13(Sin) 225
227: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
228: 6(float) ExtInst 1(GLSL.std.450) 13(Sin) 227
Store 12(inF1) 228
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
229: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
230: 6(float) ExtInst 1(GLSL.std.450) 14(Cos) 229
Store 13(inF2) 230
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
231: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
232: 6(float) ExtInst 1(GLSL.std.450) 19(Sinh) 231
233: 6(float) Load 11(inF0)
234: 6(float) Load 12(inF1)
235: 6(float) Load 13(inF2)
236: 6(float) ExtInst 1(GLSL.std.450) 49(SmoothStep) 233 234 235
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
237: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
238: 6(float) ExtInst 1(GLSL.std.450) 31(Sqrt) 237
239: 6(float) Load 11(inF0)
240: 6(float) Load 12(inF1)
241: 6(float) ExtInst 1(GLSL.std.450) 48(Step) 239 240
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
242: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
243: 6(float) ExtInst 1(GLSL.std.450) 15(Tan) 242
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
244: 6(float) Load 11(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
245: 6(float) ExtInst 1(GLSL.std.450) 21(Tanh) 244
246: 6(float) Load 11(inF0)
247: 6(float) ExtInst 1(GLSL.std.450) 3(Trunc) 246
ReturnValue 133
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
22(VertexShaderFunction1(vf1;vf1;vf1;): 6(float) Function None 18
19(inF0): 7(ptr) FunctionParameter
20(inF1): 7(ptr) FunctionParameter
21(inF2): 7(ptr) FunctionParameter
23: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
ReturnValue 133
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
34(VertexShaderFunction2(vf2;vf2;vf2;vu2;vu2;): 24(fvec2) Function None 28
29(inF0): 25(ptr) FunctionParameter
30(inF1): 25(ptr) FunctionParameter
31(inF2): 25(ptr) FunctionParameter
32(inU0): 27(ptr) FunctionParameter
33(inU1): 27(ptr) FunctionParameter
35: Label
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
252: 24(fvec2) Load 29(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
255: 253(bvec2) FUnordNotEqual 252 254
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
256: 132(bool) All 255
257: 24(fvec2) Load 29(inF0)
258: 24(fvec2) ExtInst 1(GLSL.std.450) 4(FAbs) 257
259: 24(fvec2) Load 29(inF0)
260: 24(fvec2) ExtInst 1(GLSL.std.450) 17(Acos) 259
261: 24(fvec2) Load 29(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
262: 253(bvec2) FUnordNotEqual 261 254
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
263: 132(bool) Any 262
264: 24(fvec2) Load 29(inF0)
265: 24(fvec2) ExtInst 1(GLSL.std.450) 16(Asin) 264
266: 24(fvec2) Load 29(inF0)
268: 267(ivec2) Bitcast 266
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
269: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
270: 26(ivec2) Bitcast 269
271: 26(ivec2) Load 32(inU0)
272: 24(fvec2) Bitcast 271
273: 24(fvec2) Load 29(inF0)
274: 24(fvec2) ExtInst 1(GLSL.std.450) 18(Atan) 273
275: 24(fvec2) Load 29(inF0)
276: 24(fvec2) Load 30(inF1)
277: 24(fvec2) ExtInst 1(GLSL.std.450) 25(Atan2) 275 276
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
278: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
279: 24(fvec2) ExtInst 1(GLSL.std.450) 9(Ceil) 278
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
280: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
281: 24(fvec2) Load 30(inF1)
282: 24(fvec2) Load 31(inF2)
283: 24(fvec2) ExtInst 1(GLSL.std.450) 43(FClamp) 280 281 282
284: 24(fvec2) Load 29(inF0)
285: 24(fvec2) ExtInst 1(GLSL.std.450) 14(Cos) 284
286: 24(fvec2) Load 29(inF0)
287: 24(fvec2) ExtInst 1(GLSL.std.450) 20(Cosh) 286
290: 267(ivec2) BitCount 289
291: 24(fvec2) Load 29(inF0)
292: 24(fvec2) ExtInst 1(GLSL.std.450) 12(Degrees) 291
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
293: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
294: 24(fvec2) Load 30(inF1)
295: 6(float) ExtInst 1(GLSL.std.450) 67(Distance) 293 294
296: 24(fvec2) Load 29(inF0)
297: 24(fvec2) Load 30(inF1)
298: 6(float) Dot 296 297
299: 24(fvec2) Load 29(inF0)
300: 24(fvec2) ExtInst 1(GLSL.std.450) 27(Exp) 299
301: 24(fvec2) Load 29(inF0)
302: 24(fvec2) ExtInst 1(GLSL.std.450) 29(Exp2) 301
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
303: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
304: 24(fvec2) Load 30(inF1)
305: 24(fvec2) Load 31(inF2)
306: 24(fvec2) ExtInst 1(GLSL.std.450) 70(FaceForward) 303 304 305
307: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
308: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
309: 24(fvec2) Load 29(inF0)
310: 24(fvec2) ExtInst 1(GLSL.std.450) 8(Floor) 309
311: 24(fvec2) Load 29(inF0)
312: 24(fvec2) Load 30(inF1)
313: 24(fvec2) FMod 311 312
314: 24(fvec2) Load 29(inF0)
315: 24(fvec2) ExtInst 1(GLSL.std.450) 10(Fract) 314
316: 24(fvec2) Load 29(inF0)
317: 253(bvec2) IsInf 316
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
318: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
319: 253(bvec2) IsNan 318
320: 24(fvec2) Load 29(inF0)
321: 24(fvec2) Load 30(inF1)
322: 24(fvec2) ExtInst 1(GLSL.std.450) 53(Ldexp) 320 321
323: 24(fvec2) Load 29(inF0)
324: 24(fvec2) Load 30(inF1)
325: 24(fvec2) Load 31(inF2)
326: 24(fvec2) ExtInst 1(GLSL.std.450) 46(FMix) 323 324 325
327: 24(fvec2) Load 29(inF0)
328: 6(float) ExtInst 1(GLSL.std.450) 66(Length) 327
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
329: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
330: 24(fvec2) ExtInst 1(GLSL.std.450) 28(Log) 329
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
331: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
332: 24(fvec2) ExtInst 1(GLSL.std.450) 30(Log2) 331
333: 24(fvec2) VectorTimesScalar 332 199
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
334: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
335: 24(fvec2) ExtInst 1(GLSL.std.450) 30(Log2) 334
336: 24(fvec2) Load 29(inF0)
337: 24(fvec2) Load 30(inF1)
338: 24(fvec2) ExtInst 1(GLSL.std.450) 40(FMax) 336 337
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
339: 24(fvec2) Load 29(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
340: 24(fvec2) Load 30(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
341: 24(fvec2) ExtInst 1(GLSL.std.450) 37(FMin) 339 340
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
342: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
343: 24(fvec2) ExtInst 1(GLSL.std.450) 69(Normalize) 342
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
344: 24(fvec2) Load 29(inF0)
345: 24(fvec2) Load 30(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
346: 24(fvec2) ExtInst 1(GLSL.std.450) 26(Pow) 344 345
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
347: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
348: 24(fvec2) ExtInst 1(GLSL.std.450) 11(Radians) 347
349: 24(fvec2) Load 29(inF0)
350: 24(fvec2) Load 30(inF1)
351: 24(fvec2) ExtInst 1(GLSL.std.450) 71(Reflect) 349 350
352: 24(fvec2) Load 29(inF0)
353: 24(fvec2) Load 30(inF1)
355: 24(fvec2) ExtInst 1(GLSL.std.450) 72(Refract) 352 353 354
358: 267(ivec2) BitReverse 357
359: 24(fvec2) Load 29(inF0)
360: 24(fvec2) ExtInst 1(GLSL.std.450) 2(RoundEven) 359
361: 24(fvec2) Load 29(inF0)
362: 24(fvec2) ExtInst 1(GLSL.std.450) 32(InverseSqrt) 361
363: 24(fvec2) Load 29(inF0)
364: 24(fvec2) CompositeConstruct 133 133
365: 24(fvec2) CompositeConstruct 221 221
366: 24(fvec2) ExtInst 1(GLSL.std.450) 43(FClamp) 363 364 365
367: 24(fvec2) Load 29(inF0)
368: 24(fvec2) ExtInst 1(GLSL.std.450) 6(FSign) 367
369: 24(fvec2) Load 29(inF0)
370: 24(fvec2) ExtInst 1(GLSL.std.450) 13(Sin) 369
371: 24(fvec2) Load 29(inF0)
372: 24(fvec2) ExtInst 1(GLSL.std.450) 13(Sin) 371
Store 30(inF1) 372
373: 24(fvec2) Load 29(inF0)
374: 24(fvec2) ExtInst 1(GLSL.std.450) 14(Cos) 373
Store 31(inF2) 374
375: 24(fvec2) Load 29(inF0)
376: 24(fvec2) ExtInst 1(GLSL.std.450) 19(Sinh) 375
377: 24(fvec2) Load 29(inF0)
378: 24(fvec2) Load 30(inF1)
379: 24(fvec2) Load 31(inF2)
380: 24(fvec2) ExtInst 1(GLSL.std.450) 49(SmoothStep) 377 378 379
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
381: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
382: 24(fvec2) ExtInst 1(GLSL.std.450) 31(Sqrt) 381
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
383: 24(fvec2) Load 29(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
384: 24(fvec2) Load 30(inF1)
385: 24(fvec2) ExtInst 1(GLSL.std.450) 48(Step) 383 384
386: 24(fvec2) Load 29(inF0)
387: 24(fvec2) ExtInst 1(GLSL.std.450) 15(Tan) 386
388: 24(fvec2) Load 29(inF0)
389: 24(fvec2) ExtInst 1(GLSL.std.450) 21(Tanh) 388
390: 24(fvec2) Load 29(inF0)
391: 24(fvec2) ExtInst 1(GLSL.std.450) 3(Trunc) 390
ReturnValue 392
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
46(VertexShaderFunction3(vf3;vf3;vf3;vu3;vu3;): 36(fvec3) Function None 40
41(inF0): 37(ptr) FunctionParameter
42(inF1): 37(ptr) FunctionParameter
43(inF2): 37(ptr) FunctionParameter
44(inU0): 39(ptr) FunctionParameter
45(inU1): 39(ptr) FunctionParameter
47: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
395: 36(fvec3) Load 41(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
398: 396(bvec3) FUnordNotEqual 395 397
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
399: 132(bool) All 398
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
400: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
401: 36(fvec3) ExtInst 1(GLSL.std.450) 4(FAbs) 400
402: 36(fvec3) Load 41(inF0)
403: 36(fvec3) ExtInst 1(GLSL.std.450) 17(Acos) 402
404: 36(fvec3) Load 41(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
405: 396(bvec3) FUnordNotEqual 404 397
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
406: 132(bool) Any 405
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
407: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
408: 36(fvec3) ExtInst 1(GLSL.std.450) 16(Asin) 407
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
409: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
411: 410(ivec3) Bitcast 409
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
412: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
413: 38(ivec3) Bitcast 412
414: 38(ivec3) Load 44(inU0)
415: 36(fvec3) Bitcast 414
416: 36(fvec3) Load 41(inF0)
417: 36(fvec3) ExtInst 1(GLSL.std.450) 18(Atan) 416
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
418: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
419: 36(fvec3) Load 42(inF1)
420: 36(fvec3) ExtInst 1(GLSL.std.450) 25(Atan2) 418 419
421: 36(fvec3) Load 41(inF0)
422: 36(fvec3) ExtInst 1(GLSL.std.450) 9(Ceil) 421
423: 36(fvec3) Load 41(inF0)
424: 36(fvec3) Load 42(inF1)
425: 36(fvec3) Load 43(inF2)
426: 36(fvec3) ExtInst 1(GLSL.std.450) 43(FClamp) 423 424 425
427: 36(fvec3) Load 41(inF0)
428: 36(fvec3) ExtInst 1(GLSL.std.450) 14(Cos) 427
429: 36(fvec3) Load 41(inF0)
430: 36(fvec3) ExtInst 1(GLSL.std.450) 20(Cosh) 429
433: 410(ivec3) BitCount 432
434: 36(fvec3) Load 41(inF0)
435: 36(fvec3) Load 42(inF1)
436: 36(fvec3) ExtInst 1(GLSL.std.450) 68(Cross) 434 435
437: 36(fvec3) Load 41(inF0)
438: 36(fvec3) ExtInst 1(GLSL.std.450) 12(Degrees) 437
439: 36(fvec3) Load 41(inF0)
440: 36(fvec3) Load 42(inF1)
441: 6(float) ExtInst 1(GLSL.std.450) 67(Distance) 439 440
442: 36(fvec3) Load 41(inF0)
443: 36(fvec3) Load 42(inF1)
444: 6(float) Dot 442 443
445: 36(fvec3) Load 41(inF0)
446: 36(fvec3) ExtInst 1(GLSL.std.450) 27(Exp) 445
447: 36(fvec3) Load 41(inF0)
448: 36(fvec3) ExtInst 1(GLSL.std.450) 29(Exp2) 447
449: 36(fvec3) Load 41(inF0)
450: 36(fvec3) Load 42(inF1)
451: 36(fvec3) Load 43(inF2)
452: 36(fvec3) ExtInst 1(GLSL.std.450) 70(FaceForward) 449 450 451
453: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
454: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
455: 36(fvec3) Load 41(inF0)
456: 36(fvec3) ExtInst 1(GLSL.std.450) 8(Floor) 455
457: 36(fvec3) Load 41(inF0)
458: 36(fvec3) Load 42(inF1)
459: 36(fvec3) FMod 457 458
460: 36(fvec3) Load 41(inF0)
461: 36(fvec3) ExtInst 1(GLSL.std.450) 10(Fract) 460
462: 36(fvec3) Load 41(inF0)
463: 396(bvec3) IsInf 462
464: 36(fvec3) Load 41(inF0)
465: 396(bvec3) IsNan 464
466: 36(fvec3) Load 41(inF0)
467: 36(fvec3) Load 42(inF1)
468: 36(fvec3) ExtInst 1(GLSL.std.450) 53(Ldexp) 466 467
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
469: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
470: 36(fvec3) Load 42(inF1)
471: 36(fvec3) Load 43(inF2)
472: 36(fvec3) ExtInst 1(GLSL.std.450) 46(FMix) 469 470 471
473: 36(fvec3) Load 41(inF0)
474: 6(float) ExtInst 1(GLSL.std.450) 66(Length) 473
475: 36(fvec3) Load 41(inF0)
476: 36(fvec3) ExtInst 1(GLSL.std.450) 28(Log) 475
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
477: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
478: 36(fvec3) ExtInst 1(GLSL.std.450) 30(Log2) 477
479: 36(fvec3) VectorTimesScalar 478 199
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
480: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
481: 36(fvec3) ExtInst 1(GLSL.std.450) 30(Log2) 480
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
482: 36(fvec3) Load 41(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
483: 36(fvec3) Load 42(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
484: 36(fvec3) ExtInst 1(GLSL.std.450) 40(FMax) 482 483
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
485: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
486: 36(fvec3) Load 42(inF1)
487: 36(fvec3) ExtInst 1(GLSL.std.450) 37(FMin) 485 486
488: 36(fvec3) Load 41(inF0)
489: 36(fvec3) ExtInst 1(GLSL.std.450) 69(Normalize) 488
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
490: 36(fvec3) Load 41(inF0)
491: 36(fvec3) Load 42(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
492: 36(fvec3) ExtInst 1(GLSL.std.450) 26(Pow) 490 491
493: 36(fvec3) Load 41(inF0)
494: 36(fvec3) ExtInst 1(GLSL.std.450) 11(Radians) 493
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
495: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
496: 36(fvec3) Load 42(inF1)
497: 36(fvec3) ExtInst 1(GLSL.std.450) 71(Reflect) 495 496
498: 36(fvec3) Load 41(inF0)
499: 36(fvec3) Load 42(inF1)
500: 36(fvec3) ExtInst 1(GLSL.std.450) 72(Refract) 498 499 354
502: 410(ivec3) BitReverse 501
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
503: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
504: 36(fvec3) ExtInst 1(GLSL.std.450) 2(RoundEven) 503
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
505: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
506: 36(fvec3) ExtInst 1(GLSL.std.450) 32(InverseSqrt) 505
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
507: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
508: 36(fvec3) CompositeConstruct 133 133 133
509: 36(fvec3) CompositeConstruct 221 221 221
510: 36(fvec3) ExtInst 1(GLSL.std.450) 43(FClamp) 507 508 509
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
511: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
512: 36(fvec3) ExtInst 1(GLSL.std.450) 6(FSign) 511
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
513: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
514: 36(fvec3) ExtInst 1(GLSL.std.450) 13(Sin) 513
515: 36(fvec3) Load 41(inF0)
516: 36(fvec3) ExtInst 1(GLSL.std.450) 13(Sin) 515
Store 42(inF1) 516
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
517: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
518: 36(fvec3) ExtInst 1(GLSL.std.450) 14(Cos) 517
Store 43(inF2) 518
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
519: 36(fvec3) Load 41(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
520: 36(fvec3) ExtInst 1(GLSL.std.450) 19(Sinh) 519
521: 36(fvec3) Load 41(inF0)
522: 36(fvec3) Load 42(inF1)
523: 36(fvec3) Load 43(inF2)
524: 36(fvec3) ExtInst 1(GLSL.std.450) 49(SmoothStep) 521 522 523
525: 36(fvec3) Load 41(inF0)
526: 36(fvec3) ExtInst 1(GLSL.std.450) 31(Sqrt) 525
527: 36(fvec3) Load 41(inF0)
528: 36(fvec3) Load 42(inF1)
529: 36(fvec3) ExtInst 1(GLSL.std.450) 48(Step) 527 528
530: 36(fvec3) Load 41(inF0)
531: 36(fvec3) ExtInst 1(GLSL.std.450) 15(Tan) 530
532: 36(fvec3) Load 41(inF0)
533: 36(fvec3) ExtInst 1(GLSL.std.450) 21(Tanh) 532
534: 36(fvec3) Load 41(inF0)
535: 36(fvec3) ExtInst 1(GLSL.std.450) 3(Trunc) 534
ReturnValue 537
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
58(VertexShaderFunction4(vf4;vf4;vf4;vu4;vu4;): 48(fvec4) Function None 52
53(inF0): 49(ptr) FunctionParameter
54(inF1): 49(ptr) FunctionParameter
55(inF2): 49(ptr) FunctionParameter
56(inU0): 51(ptr) FunctionParameter
57(inU1): 51(ptr) FunctionParameter
59: Label
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
540: 48(fvec4) Load 53(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
543: 541(bvec4) FUnordNotEqual 540 542
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
544: 132(bool) All 543
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
545: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
546: 48(fvec4) ExtInst 1(GLSL.std.450) 4(FAbs) 545
547: 48(fvec4) Load 53(inF0)
548: 48(fvec4) ExtInst 1(GLSL.std.450) 17(Acos) 547
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
549: 48(fvec4) Load 53(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
550: 541(bvec4) FUnordNotEqual 549 542
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
551: 132(bool) Any 550
552: 48(fvec4) Load 53(inF0)
553: 48(fvec4) ExtInst 1(GLSL.std.450) 16(Asin) 552
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
554: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
556: 555(ivec4) Bitcast 554
557: 48(fvec4) Load 53(inF0)
558: 50(ivec4) Bitcast 557
559: 50(ivec4) Load 56(inU0)
560: 48(fvec4) Bitcast 559
561: 48(fvec4) Load 53(inF0)
562: 48(fvec4) ExtInst 1(GLSL.std.450) 18(Atan) 561
563: 48(fvec4) Load 53(inF0)
564: 48(fvec4) Load 54(inF1)
565: 48(fvec4) ExtInst 1(GLSL.std.450) 25(Atan2) 563 564
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
566: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
567: 48(fvec4) ExtInst 1(GLSL.std.450) 9(Ceil) 566
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
568: 48(fvec4) Load 53(inF0)
569: 48(fvec4) Load 54(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
570: 48(fvec4) Load 55(inF2)
571: 48(fvec4) ExtInst 1(GLSL.std.450) 43(FClamp) 568 569 570
572: 48(fvec4) Load 53(inF0)
573: 48(fvec4) ExtInst 1(GLSL.std.450) 14(Cos) 572
574: 48(fvec4) Load 53(inF0)
575: 48(fvec4) ExtInst 1(GLSL.std.450) 20(Cosh) 574
577: 555(ivec4) BitCount 576
578: 48(fvec4) Load 53(inF0)
579: 48(fvec4) ExtInst 1(GLSL.std.450) 12(Degrees) 578
580: 48(fvec4) Load 53(inF0)
581: 48(fvec4) Load 54(inF1)
582: 6(float) ExtInst 1(GLSL.std.450) 67(Distance) 580 581
583: 48(fvec4) Load 53(inF0)
584: 48(fvec4) Load 54(inF1)
585: 6(float) Dot 583 584
587: 7(ptr) AccessChain 53(inF0) 586
588: 6(float) Load 587
589: 7(ptr) AccessChain 54(inF1) 586
590: 6(float) Load 589
591: 6(float) FMul 588 590
593: 7(ptr) AccessChain 53(inF0) 592
594: 6(float) Load 593
596: 7(ptr) AccessChain 54(inF1) 595
597: 6(float) Load 596
598: 48(fvec4) CompositeConstruct 221 591 594 597
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
599: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
600: 48(fvec4) ExtInst 1(GLSL.std.450) 27(Exp) 599
601: 48(fvec4) Load 53(inF0)
602: 48(fvec4) ExtInst 1(GLSL.std.450) 29(Exp2) 601
603: 48(fvec4) Load 53(inF0)
604: 48(fvec4) Load 54(inF1)
605: 48(fvec4) Load 55(inF2)
606: 48(fvec4) ExtInst 1(GLSL.std.450) 70(FaceForward) 603 604 605
607: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
608: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
609: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
610: 48(fvec4) ExtInst 1(GLSL.std.450) 8(Floor) 609
611: 48(fvec4) Load 53(inF0)
612: 48(fvec4) Load 54(inF1)
613: 48(fvec4) FMod 611 612
614: 48(fvec4) Load 53(inF0)
615: 48(fvec4) ExtInst 1(GLSL.std.450) 10(Fract) 614
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
616: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
617: 541(bvec4) IsInf 616
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
618: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
619: 541(bvec4) IsNan 618
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
620: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
621: 48(fvec4) Load 54(inF1)
622: 48(fvec4) ExtInst 1(GLSL.std.450) 53(Ldexp) 620 621
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
623: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
624: 48(fvec4) Load 54(inF1)
625: 48(fvec4) Load 55(inF2)
626: 48(fvec4) ExtInst 1(GLSL.std.450) 46(FMix) 623 624 625
627: 48(fvec4) Load 53(inF0)
628: 6(float) ExtInst 1(GLSL.std.450) 66(Length) 627
629: 48(fvec4) Load 53(inF0)
630: 48(fvec4) ExtInst 1(GLSL.std.450) 28(Log) 629
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
631: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
632: 48(fvec4) ExtInst 1(GLSL.std.450) 30(Log2) 631
633: 48(fvec4) VectorTimesScalar 632 199
634: 48(fvec4) Load 53(inF0)
635: 48(fvec4) ExtInst 1(GLSL.std.450) 30(Log2) 634
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
636: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
637: 48(fvec4) Load 54(inF1)
638: 48(fvec4) ExtInst 1(GLSL.std.450) 40(FMax) 636 637
639: 48(fvec4) Load 53(inF0)
640: 48(fvec4) Load 54(inF1)
641: 48(fvec4) ExtInst 1(GLSL.std.450) 37(FMin) 639 640
642: 48(fvec4) Load 53(inF0)
643: 48(fvec4) ExtInst 1(GLSL.std.450) 69(Normalize) 642
644: 48(fvec4) Load 53(inF0)
645: 48(fvec4) Load 54(inF1)
646: 48(fvec4) ExtInst 1(GLSL.std.450) 26(Pow) 644 645
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
647: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
648: 48(fvec4) ExtInst 1(GLSL.std.450) 11(Radians) 647
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
649: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
650: 48(fvec4) Load 54(inF1)
651: 48(fvec4) ExtInst 1(GLSL.std.450) 71(Reflect) 649 650
652: 48(fvec4) Load 53(inF0)
653: 48(fvec4) Load 54(inF1)
654: 48(fvec4) ExtInst 1(GLSL.std.450) 72(Refract) 652 653 354
657: 555(ivec4) BitReverse 656
658: 48(fvec4) Load 53(inF0)
659: 48(fvec4) ExtInst 1(GLSL.std.450) 2(RoundEven) 658
660: 48(fvec4) Load 53(inF0)
661: 48(fvec4) ExtInst 1(GLSL.std.450) 32(InverseSqrt) 660
662: 48(fvec4) Load 53(inF0)
663: 48(fvec4) CompositeConstruct 133 133 133 133
664: 48(fvec4) CompositeConstruct 221 221 221 221
665: 48(fvec4) ExtInst 1(GLSL.std.450) 43(FClamp) 662 663 664
666: 48(fvec4) Load 53(inF0)
667: 48(fvec4) ExtInst 1(GLSL.std.450) 6(FSign) 666
668: 48(fvec4) Load 53(inF0)
669: 48(fvec4) ExtInst 1(GLSL.std.450) 13(Sin) 668
670: 48(fvec4) Load 53(inF0)
671: 48(fvec4) ExtInst 1(GLSL.std.450) 13(Sin) 670
Store 54(inF1) 671
672: 48(fvec4) Load 53(inF0)
673: 48(fvec4) ExtInst 1(GLSL.std.450) 14(Cos) 672
Store 55(inF2) 673
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
674: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
675: 48(fvec4) ExtInst 1(GLSL.std.450) 19(Sinh) 674
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
676: 48(fvec4) Load 53(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
677: 48(fvec4) Load 54(inF1)
678: 48(fvec4) Load 55(inF2)
679: 48(fvec4) ExtInst 1(GLSL.std.450) 49(SmoothStep) 676 677 678
680: 48(fvec4) Load 53(inF0)
681: 48(fvec4) ExtInst 1(GLSL.std.450) 31(Sqrt) 680
682: 48(fvec4) Load 53(inF0)
683: 48(fvec4) Load 54(inF1)
684: 48(fvec4) ExtInst 1(GLSL.std.450) 48(Step) 682 683
685: 48(fvec4) Load 53(inF0)
686: 48(fvec4) ExtInst 1(GLSL.std.450) 15(Tan) 685
687: 48(fvec4) Load 53(inF0)
688: 48(fvec4) ExtInst 1(GLSL.std.450) 21(Tanh) 687
689: 48(fvec4) Load 53(inF0)
690: 48(fvec4) ExtInst 1(GLSL.std.450) 3(Trunc) 689
ReturnValue 692
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
66(VertexShaderFunction2x2(mf22;mf22;mf22;): 60 Function None 62
63(inF0): 61(ptr) FunctionParameter
64(inF1): 61(ptr) FunctionParameter
65(inF2): 61(ptr) FunctionParameter
67: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
695: 60 Load 63(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
697: 696 FUnordNotEqual 695 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
698: 132(bool) All 697
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
699: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
700: 60 ExtInst 1(GLSL.std.450) 4(FAbs) 699
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
701: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
702: 60 ExtInst 1(GLSL.std.450) 17(Acos) 701
703: 60 Load 63(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
704: 696 FUnordNotEqual 703 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
705: 132(bool) Any 704
706: 60 Load 63(inF0)
707: 60 ExtInst 1(GLSL.std.450) 16(Asin) 706
708: 60 Load 63(inF0)
709: 60 ExtInst 1(GLSL.std.450) 18(Atan) 708
710: 60 Load 63(inF0)
711: 60 Load 64(inF1)
712: 60 ExtInst 1(GLSL.std.450) 25(Atan2) 710 711
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
713: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
714: 60 ExtInst 1(GLSL.std.450) 9(Ceil) 713
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
715: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
716: 60 Load 64(inF1)
717: 60 Load 65(inF2)
718: 60 ExtInst 1(GLSL.std.450) 43(FClamp) 715 716 717
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
719: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
720: 60 ExtInst 1(GLSL.std.450) 14(Cos) 719
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
721: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
722: 60 ExtInst 1(GLSL.std.450) 20(Cosh) 721
723: 60 Load 63(inF0)
724: 60 ExtInst 1(GLSL.std.450) 12(Degrees) 723
725: 60 Load 63(inF0)
726: 6(float) ExtInst 1(GLSL.std.450) 33(Determinant) 725
727: 60 Load 63(inF0)
728: 60 ExtInst 1(GLSL.std.450) 27(Exp) 727
729: 60 Load 63(inF0)
730: 60 ExtInst 1(GLSL.std.450) 29(Exp2) 729
731: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
732: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
733: 60 Load 63(inF0)
734: 60 ExtInst 1(GLSL.std.450) 8(Floor) 733
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
735: 60 Load 63(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
736: 60 Load 64(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
737: 24(fvec2) CompositeExtract 735 0
738: 24(fvec2) CompositeExtract 736 0
739: 24(fvec2) FMod 737 738
740: 24(fvec2) CompositeExtract 735 1
741: 24(fvec2) CompositeExtract 736 1
742: 24(fvec2) FMod 740 741
743: 60 CompositeConstruct 739 742
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
744: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
745: 60 ExtInst 1(GLSL.std.450) 10(Fract) 744
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
746: 60 Load 63(inF0)
747: 60 Load 64(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
748: 60 ExtInst 1(GLSL.std.450) 53(Ldexp) 746 747
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
749: 60 Load 63(inF0)
750: 60 Load 64(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
751: 60 Load 65(inF2)
752: 60 ExtInst 1(GLSL.std.450) 46(FMix) 749 750 751
753: 60 Load 63(inF0)
754: 60 ExtInst 1(GLSL.std.450) 28(Log) 753
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
755: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
756: 60 ExtInst 1(GLSL.std.450) 30(Log2) 755
757: 60 MatrixTimesScalar 756 199
758: 60 Load 63(inF0)
759: 60 ExtInst 1(GLSL.std.450) 30(Log2) 758
760: 60 Load 63(inF0)
761: 60 Load 64(inF1)
762: 60 ExtInst 1(GLSL.std.450) 40(FMax) 760 761
763: 60 Load 63(inF0)
764: 60 Load 64(inF1)
765: 60 ExtInst 1(GLSL.std.450) 37(FMin) 763 764
766: 60 Load 63(inF0)
767: 60 Load 64(inF1)
768: 60 ExtInst 1(GLSL.std.450) 26(Pow) 766 767
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
769: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
770: 60 ExtInst 1(GLSL.std.450) 11(Radians) 769
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
771: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
772: 60 ExtInst 1(GLSL.std.450) 2(RoundEven) 771
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
773: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
774: 60 ExtInst 1(GLSL.std.450) 32(InverseSqrt) 773
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
775: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
776: 24(fvec2) CompositeConstruct 133 133
777: 24(fvec2) CompositeConstruct 221 221
778: 60 ExtInst 1(GLSL.std.450) 43(FClamp) 775 776 777
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
779: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
780: 60 ExtInst 1(GLSL.std.450) 6(FSign) 779
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
781: 60 Load 63(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
782: 60 ExtInst 1(GLSL.std.450) 13(Sin) 781
783: 60 Load 63(inF0)
784: 60 ExtInst 1(GLSL.std.450) 13(Sin) 783
Store 64(inF1) 784
785: 60 Load 63(inF0)
786: 60 ExtInst 1(GLSL.std.450) 14(Cos) 785
Store 65(inF2) 786
787: 60 Load 63(inF0)
788: 60 ExtInst 1(GLSL.std.450) 19(Sinh) 787
789: 60 Load 63(inF0)
790: 60 Load 64(inF1)
791: 60 Load 65(inF2)
792: 60 ExtInst 1(GLSL.std.450) 49(SmoothStep) 789 790 791
793: 60 Load 63(inF0)
794: 60 ExtInst 1(GLSL.std.450) 31(Sqrt) 793
795: 60 Load 63(inF0)
796: 60 Load 64(inF1)
797: 60 ExtInst 1(GLSL.std.450) 48(Step) 795 796
798: 60 Load 63(inF0)
799: 60 ExtInst 1(GLSL.std.450) 15(Tan) 798
800: 60 Load 63(inF0)
801: 60 ExtInst 1(GLSL.std.450) 21(Tanh) 800
802: 60 Load 63(inF0)
803: 60 Transpose 802
804: 60 Load 63(inF0)
805: 60 ExtInst 1(GLSL.std.450) 3(Trunc) 804
ReturnValue 807
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
74(VertexShaderFunction3x3(mf33;mf33;mf33;): 68 Function None 70
71(inF0): 69(ptr) FunctionParameter
72(inF1): 69(ptr) FunctionParameter
73(inF2): 69(ptr) FunctionParameter
75: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
810: 68 Load 71(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
812: 811 FUnordNotEqual 810 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
813: 132(bool) All 812
814: 68 Load 71(inF0)
815: 68 ExtInst 1(GLSL.std.450) 4(FAbs) 814
816: 68 Load 71(inF0)
817: 68 ExtInst 1(GLSL.std.450) 17(Acos) 816
818: 68 Load 71(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
819: 811 FUnordNotEqual 818 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
820: 132(bool) Any 819
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
821: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
822: 68 ExtInst 1(GLSL.std.450) 16(Asin) 821
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
823: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
824: 68 ExtInst 1(GLSL.std.450) 18(Atan) 823
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
825: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
826: 68 Load 72(inF1)
827: 68 ExtInst 1(GLSL.std.450) 25(Atan2) 825 826
828: 68 Load 71(inF0)
829: 68 ExtInst 1(GLSL.std.450) 9(Ceil) 828
830: 68 Load 71(inF0)
831: 68 Load 72(inF1)
832: 68 Load 73(inF2)
833: 68 ExtInst 1(GLSL.std.450) 43(FClamp) 830 831 832
834: 68 Load 71(inF0)
835: 68 ExtInst 1(GLSL.std.450) 14(Cos) 834
836: 68 Load 71(inF0)
837: 68 ExtInst 1(GLSL.std.450) 20(Cosh) 836
838: 68 Load 71(inF0)
839: 68 ExtInst 1(GLSL.std.450) 12(Degrees) 838
840: 68 Load 71(inF0)
841: 6(float) ExtInst 1(GLSL.std.450) 33(Determinant) 840
842: 68 Load 71(inF0)
843: 68 ExtInst 1(GLSL.std.450) 27(Exp) 842
844: 68 Load 71(inF0)
845: 68 ExtInst 1(GLSL.std.450) 29(Exp2) 844
846: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
847: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
848: 68 Load 71(inF0)
849: 68 ExtInst 1(GLSL.std.450) 8(Floor) 848
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
850: 68 Load 71(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
851: 68 Load 72(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
852: 36(fvec3) CompositeExtract 850 0
853: 36(fvec3) CompositeExtract 851 0
854: 36(fvec3) FMod 852 853
855: 36(fvec3) CompositeExtract 850 1
856: 36(fvec3) CompositeExtract 851 1
857: 36(fvec3) FMod 855 856
858: 36(fvec3) CompositeExtract 850 2
859: 36(fvec3) CompositeExtract 851 2
860: 36(fvec3) FMod 858 859
861: 68 CompositeConstruct 854 857 860
862: 68 Load 71(inF0)
863: 68 ExtInst 1(GLSL.std.450) 10(Fract) 862
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
864: 68 Load 71(inF0)
865: 68 Load 72(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
866: 68 ExtInst 1(GLSL.std.450) 53(Ldexp) 864 865
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
867: 68 Load 71(inF0)
868: 68 Load 72(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
869: 68 Load 73(inF2)
870: 68 ExtInst 1(GLSL.std.450) 46(FMix) 867 868 869
871: 68 Load 71(inF0)
872: 68 ExtInst 1(GLSL.std.450) 28(Log) 871
873: 68 Load 71(inF0)
874: 68 ExtInst 1(GLSL.std.450) 30(Log2) 873
875: 68 MatrixTimesScalar 874 199
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
876: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
877: 68 ExtInst 1(GLSL.std.450) 30(Log2) 876
878: 68 Load 71(inF0)
879: 68 Load 72(inF1)
880: 68 ExtInst 1(GLSL.std.450) 40(FMax) 878 879
881: 68 Load 71(inF0)
882: 68 Load 72(inF1)
883: 68 ExtInst 1(GLSL.std.450) 37(FMin) 881 882
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
884: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
885: 68 Load 72(inF1)
886: 68 ExtInst 1(GLSL.std.450) 26(Pow) 884 885
887: 68 Load 71(inF0)
888: 68 ExtInst 1(GLSL.std.450) 11(Radians) 887
889: 68 Load 71(inF0)
890: 68 ExtInst 1(GLSL.std.450) 2(RoundEven) 889
891: 68 Load 71(inF0)
892: 68 ExtInst 1(GLSL.std.450) 32(InverseSqrt) 891
893: 68 Load 71(inF0)
894: 36(fvec3) CompositeConstruct 133 133 133
895: 36(fvec3) CompositeConstruct 221 221 221
896: 68 ExtInst 1(GLSL.std.450) 43(FClamp) 893 894 895
897: 68 Load 71(inF0)
898: 68 ExtInst 1(GLSL.std.450) 6(FSign) 897
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
899: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
900: 68 ExtInst 1(GLSL.std.450) 13(Sin) 899
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
901: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
902: 68 ExtInst 1(GLSL.std.450) 13(Sin) 901
Store 72(inF1) 902
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
903: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
904: 68 ExtInst 1(GLSL.std.450) 14(Cos) 903
Store 73(inF2) 904
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
905: 68 Load 71(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
906: 68 ExtInst 1(GLSL.std.450) 19(Sinh) 905
907: 68 Load 71(inF0)
908: 68 Load 72(inF1)
909: 68 Load 73(inF2)
910: 68 ExtInst 1(GLSL.std.450) 49(SmoothStep) 907 908 909
911: 68 Load 71(inF0)
912: 68 ExtInst 1(GLSL.std.450) 31(Sqrt) 911
913: 68 Load 71(inF0)
914: 68 Load 72(inF1)
915: 68 ExtInst 1(GLSL.std.450) 48(Step) 913 914
916: 68 Load 71(inF0)
917: 68 ExtInst 1(GLSL.std.450) 15(Tan) 916
918: 68 Load 71(inF0)
919: 68 ExtInst 1(GLSL.std.450) 21(Tanh) 918
920: 68 Load 71(inF0)
921: 68 Transpose 920
922: 68 Load 71(inF0)
923: 68 ExtInst 1(GLSL.std.450) 3(Trunc) 922
ReturnValue 925
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
82(VertexShaderFunction4x4(mf44;mf44;mf44;): 76 Function None 78
79(inF0): 77(ptr) FunctionParameter
80(inF1): 77(ptr) FunctionParameter
81(inF2): 77(ptr) FunctionParameter
83: Label
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
928: 76 Load 79(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
930: 929 FUnordNotEqual 928 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
931: 132(bool) All 930
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
932: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
933: 76 ExtInst 1(GLSL.std.450) 4(FAbs) 932
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
934: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
935: 76 ExtInst 1(GLSL.std.450) 17(Acos) 934
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
936: 76 Load 79(inF0)
Update test results to expect OpFUnordNotEqual
2020-06-05 13:30:05 +00:00
937: 929 FUnordNotEqual 936 133
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
938: 132(bool) Any 937
939: 76 Load 79(inF0)
940: 76 ExtInst 1(GLSL.std.450) 16(Asin) 939
941: 76 Load 79(inF0)
942: 76 ExtInst 1(GLSL.std.450) 18(Atan) 941
943: 76 Load 79(inF0)
944: 76 Load 80(inF1)
945: 76 ExtInst 1(GLSL.std.450) 25(Atan2) 943 944
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
946: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
947: 76 ExtInst 1(GLSL.std.450) 9(Ceil) 946
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
948: 76 Load 79(inF0)
949: 76 Load 80(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
950: 76 Load 81(inF2)
951: 76 ExtInst 1(GLSL.std.450) 43(FClamp) 948 949 950
952: 76 Load 79(inF0)
953: 76 ExtInst 1(GLSL.std.450) 14(Cos) 952
954: 76 Load 79(inF0)
955: 76 ExtInst 1(GLSL.std.450) 20(Cosh) 954
956: 76 Load 79(inF0)
957: 76 ExtInst 1(GLSL.std.450) 12(Degrees) 956
958: 76 Load 79(inF0)
959: 6(float) ExtInst 1(GLSL.std.450) 33(Determinant) 958
960: 76 Load 79(inF0)
961: 76 ExtInst 1(GLSL.std.450) 27(Exp) 960
962: 76 Load 79(inF0)
963: 76 ExtInst 1(GLSL.std.450) 29(Exp2) 962
964: 146(int) ExtInst 1(GLSL.std.450) 74(FindSMsb) 167
965: 146(int) ExtInst 1(GLSL.std.450) 73(FindILsb) 167
966: 76 Load 79(inF0)
967: 76 ExtInst 1(GLSL.std.450) 8(Floor) 966
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
968: 76 Load 79(inF0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
969: 76 Load 80(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
970: 48(fvec4) CompositeExtract 968 0
971: 48(fvec4) CompositeExtract 969 0
972: 48(fvec4) FMod 970 971
973: 48(fvec4) CompositeExtract 968 1
974: 48(fvec4) CompositeExtract 969 1
975: 48(fvec4) FMod 973 974
976: 48(fvec4) CompositeExtract 968 2
977: 48(fvec4) CompositeExtract 969 2
978: 48(fvec4) FMod 976 977
979: 48(fvec4) CompositeExtract 968 3
980: 48(fvec4) CompositeExtract 969 3
981: 48(fvec4) FMod 979 980
982: 76 CompositeConstruct 972 975 978 981
983: 76 Load 79(inF0)
984: 76 ExtInst 1(GLSL.std.450) 10(Fract) 983
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
985: 76 Load 79(inF0)
986: 76 Load 80(inF1)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
987: 76 ExtInst 1(GLSL.std.450) 53(Ldexp) 985 986
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
988: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
989: 76 Load 80(inF1)
990: 76 Load 81(inF2)
991: 76 ExtInst 1(GLSL.std.450) 46(FMix) 988 989 990
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
992: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
993: 76 ExtInst 1(GLSL.std.450) 28(Log) 992
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
994: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
995: 76 ExtInst 1(GLSL.std.450) 30(Log2) 994
996: 76 MatrixTimesScalar 995 199
997: 76 Load 79(inF0)
998: 76 ExtInst 1(GLSL.std.450) 30(Log2) 997
999: 76 Load 79(inF0)
1000: 76 Load 80(inF1)
1001: 76 ExtInst 1(GLSL.std.450) 40(FMax) 999 1000
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1002: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1003: 76 Load 80(inF1)
1004: 76 ExtInst 1(GLSL.std.450) 37(FMin) 1002 1003
1005: 76 Load 79(inF0)
1006: 76 Load 80(inF1)
1007: 76 ExtInst 1(GLSL.std.450) 26(Pow) 1005 1006
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1008: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1009: 76 ExtInst 1(GLSL.std.450) 11(Radians) 1008
1010: 76 Load 79(inF0)
1011: 76 ExtInst 1(GLSL.std.450) 2(RoundEven) 1010
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
1012: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1013: 76 ExtInst 1(GLSL.std.450) 32(InverseSqrt) 1012
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
1014: 76 Load 79(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1015: 48(fvec4) CompositeConstruct 133 133 133 133
1016: 48(fvec4) CompositeConstruct 221 221 221 221
1017: 76 ExtInst 1(GLSL.std.450) 43(FClamp) 1014 1015 1016
1018: 76 Load 79(inF0)
1019: 76 ExtInst 1(GLSL.std.450) 6(FSign) 1018
1020: 76 Load 79(inF0)
1021: 76 ExtInst 1(GLSL.std.450) 13(Sin) 1020
1022: 76 Load 79(inF0)
1023: 76 ExtInst 1(GLSL.std.450) 13(Sin) 1022
Store 80(inF1) 1023
1024: 76 Load 79(inF0)
1025: 76 ExtInst 1(GLSL.std.450) 14(Cos) 1024
Store 81(inF2) 1025
1026: 76 Load 79(inF0)
1027: 76 ExtInst 1(GLSL.std.450) 19(Sinh) 1026
1028: 76 Load 79(inF0)
1029: 76 Load 80(inF1)
1030: 76 Load 81(inF2)
1031: 76 ExtInst 1(GLSL.std.450) 49(SmoothStep) 1028 1029 1030
1032: 76 Load 79(inF0)
1033: 76 ExtInst 1(GLSL.std.450) 31(Sqrt) 1032
1034: 76 Load 79(inF0)
1035: 76 Load 80(inF1)
1036: 76 ExtInst 1(GLSL.std.450) 48(Step) 1034 1035
1037: 76 Load 79(inF0)
1038: 76 ExtInst 1(GLSL.std.450) 15(Tan) 1037
1039: 76 Load 79(inF0)
1040: 76 ExtInst 1(GLSL.std.450) 21(Tanh) 1039
1041: 76 Load 79(inF0)
1042: 76 Transpose 1041
1043: 76 Load 79(inF0)
1044: 76 ExtInst 1(GLSL.std.450) 3(Trunc) 1043
ReturnValue 1046
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
FunctionEnd
91(TestGenMul2(f1;f1;vf2;vf2;mf22;mf22;): 2 Function None 84
85(inF0): 7(ptr) FunctionParameter
86(inF1): 7(ptr) FunctionParameter
87(inFV0): 25(ptr) FunctionParameter
88(inFV1): 25(ptr) FunctionParameter
89(inFM0): 61(ptr) FunctionParameter
90(inFM1): 61(ptr) FunctionParameter
92: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1049(r0): 7(ptr) Variable Function
1053(r1): 25(ptr) Variable Function
1057(r2): 25(ptr) Variable Function
1061(r3): 7(ptr) Variable Function
1065(r4): 25(ptr) Variable Function
1069(r5): 25(ptr) Variable Function
1073(r6): 61(ptr) Variable Function
1077(r7): 61(ptr) Variable Function
1081(r8): 61(ptr) Variable Function
1050: 6(float) Load 86(inF1)
1051: 6(float) Load 85(inF0)
1052: 6(float) FMul 1050 1051
Store 1049(r0) 1052
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1054: 6(float) Load 85(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1055: 24(fvec2) Load 87(inFV0)
1056: 24(fvec2) VectorTimesScalar 1055 1054
Store 1053(r1) 1056
1058: 24(fvec2) Load 87(inFV0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1059: 6(float) Load 85(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1060: 24(fvec2) VectorTimesScalar 1058 1059
Store 1057(r2) 1060
1062: 24(fvec2) Load 87(inFV0)
1063: 24(fvec2) Load 88(inFV1)
1064: 6(float) Dot 1062 1063
Store 1061(r3) 1064
1066: 24(fvec2) Load 87(inFV0)
1067: 60 Load 89(inFM0)
1068: 24(fvec2) VectorTimesMatrix 1066 1067
Store 1065(r4) 1068
1070: 60 Load 89(inFM0)
1071: 24(fvec2) Load 87(inFV0)
1072: 24(fvec2) MatrixTimesVector 1070 1071
Store 1069(r5) 1072
1074: 6(float) Load 85(inF0)
1075: 60 Load 89(inFM0)
1076: 60 MatrixTimesScalar 1075 1074
Store 1073(r6) 1076
1078: 60 Load 89(inFM0)
1079: 6(float) Load 85(inF0)
1080: 60 MatrixTimesScalar 1078 1079
Store 1077(r7) 1080
1082: 60 Load 90(inFM1)
1083: 60 Load 89(inFM0)
1084: 60 MatrixTimesMatrix 1082 1083
Store 1081(r8) 1084
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
Return
FunctionEnd
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
100(TestGenMul3(f1;f1;vf3;vf3;mf33;mf33;): 2 Function None 93
94(inF0): 7(ptr) FunctionParameter
95(inF1): 7(ptr) FunctionParameter
96(inFV0): 37(ptr) FunctionParameter
97(inFV1): 37(ptr) FunctionParameter
98(inFM0): 69(ptr) FunctionParameter
99(inFM1): 69(ptr) FunctionParameter
101: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1085(r0): 7(ptr) Variable Function
1089(r1): 37(ptr) Variable Function
1093(r2): 37(ptr) Variable Function
1097(r3): 7(ptr) Variable Function
1101(r4): 37(ptr) Variable Function
1105(r5): 37(ptr) Variable Function
1109(r6): 69(ptr) Variable Function
1113(r7): 69(ptr) Variable Function
1117(r8): 69(ptr) Variable Function
1086: 6(float) Load 95(inF1)
1087: 6(float) Load 94(inF0)
1088: 6(float) FMul 1086 1087
Store 1085(r0) 1088
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1090: 6(float) Load 94(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1091: 36(fvec3) Load 96(inFV0)
1092: 36(fvec3) VectorTimesScalar 1091 1090
Store 1089(r1) 1092
1094: 36(fvec3) Load 96(inFV0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1095: 6(float) Load 94(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1096: 36(fvec3) VectorTimesScalar 1094 1095
Store 1093(r2) 1096
1098: 36(fvec3) Load 96(inFV0)
1099: 36(fvec3) Load 97(inFV1)
1100: 6(float) Dot 1098 1099
Store 1097(r3) 1100
1102: 36(fvec3) Load 96(inFV0)
1103: 68 Load 98(inFM0)
1104: 36(fvec3) VectorTimesMatrix 1102 1103
Store 1101(r4) 1104
1106: 68 Load 98(inFM0)
1107: 36(fvec3) Load 96(inFV0)
1108: 36(fvec3) MatrixTimesVector 1106 1107
Store 1105(r5) 1108
1110: 6(float) Load 94(inF0)
1111: 68 Load 98(inFM0)
1112: 68 MatrixTimesScalar 1111 1110
Store 1109(r6) 1112
1114: 68 Load 98(inFM0)
1115: 6(float) Load 94(inF0)
1116: 68 MatrixTimesScalar 1114 1115
Store 1113(r7) 1116
1118: 68 Load 99(inFM1)
1119: 68 Load 98(inFM0)
1120: 68 MatrixTimesMatrix 1118 1119
Store 1117(r8) 1120
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
Return
FunctionEnd
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
109(TestGenMul4(f1;f1;vf4;vf4;mf44;mf44;): 2 Function None 102
103(inF0): 7(ptr) FunctionParameter
104(inF1): 7(ptr) FunctionParameter
105(inFV0): 49(ptr) FunctionParameter
106(inFV1): 49(ptr) FunctionParameter
107(inFM0): 77(ptr) FunctionParameter
108(inFM1): 77(ptr) FunctionParameter
110: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1121(r0): 7(ptr) Variable Function
1125(r1): 49(ptr) Variable Function
1129(r2): 49(ptr) Variable Function
1133(r3): 7(ptr) Variable Function
1137(r4): 49(ptr) Variable Function
1141(r5): 49(ptr) Variable Function
1145(r6): 77(ptr) Variable Function
1149(r7): 77(ptr) Variable Function
1153(r8): 77(ptr) Variable Function
1122: 6(float) Load 104(inF1)
1123: 6(float) Load 103(inF0)
1124: 6(float) FMul 1122 1123
Store 1121(r0) 1124
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1126: 6(float) Load 103(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1127: 48(fvec4) Load 105(inFV0)
1128: 48(fvec4) VectorTimesScalar 1127 1126
Store 1125(r1) 1128
1130: 48(fvec4) Load 105(inFV0)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1131: 6(float) Load 103(inF0)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1132: 48(fvec4) VectorTimesScalar 1130 1131
Store 1129(r2) 1132
1134: 48(fvec4) Load 105(inFV0)
1135: 48(fvec4) Load 106(inFV1)
1136: 6(float) Dot 1134 1135
Store 1133(r3) 1136
1138: 48(fvec4) Load 105(inFV0)
1139: 76 Load 107(inFM0)
1140: 48(fvec4) VectorTimesMatrix 1138 1139
Store 1137(r4) 1140
1142: 76 Load 107(inFM0)
1143: 48(fvec4) Load 105(inFV0)
1144: 48(fvec4) MatrixTimesVector 1142 1143
Store 1141(r5) 1144
1146: 6(float) Load 103(inF0)
1147: 76 Load 107(inFM0)
1148: 76 MatrixTimesScalar 1147 1146
Store 1145(r6) 1148
1150: 76 Load 107(inFM0)
1151: 6(float) Load 103(inF0)
1152: 76 MatrixTimesScalar 1150 1151
Store 1149(r7) 1152
1154: 76 Load 108(inFM1)
1155: 76 Load 107(inFM0)
1156: 76 MatrixTimesMatrix 1154 1155
Store 1153(r8) 1156
Add lerp, fix sincos return type, ret type tests, non-square mats, HLSL->AST tx semantics
2016-06-20 19:26:59 +00:00
Return
FunctionEnd
WIP: HLSL: Treat HLSL rows as GLSL columns. WIP: HLSL: EOpGenMul arg reversal
2016-08-24 20:36:13 +00:00
129(TestGenMulNxM(f1;f1;vf2;vf3;mf23;mf32;mf33;mf34;mf24;): 2 Function None 119
HLSL: Change intrinsics tests to not use overloaded entry points
2016-06-30 20:13:03 +00:00
120(inF0): 7(ptr) FunctionParameter
121(inF1): 7(ptr) FunctionParameter
122(inFV2): 25(ptr) FunctionParameter
123(inFV3): 37(ptr) FunctionParameter
124(inFM2x3): 112(ptr) FunctionParameter
125(inFM3x2): 114(ptr) FunctionParameter
126(inFM3x3): 69(ptr) FunctionParameter
127(inFM3x4): 116(ptr) FunctionParameter
128(inFM2x4): 118(ptr) FunctionParameter
130: Label
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1157(r00): 7(ptr) Variable Function
1161(r01): 25(ptr) Variable Function
1165(r02): 37(ptr) Variable Function
1169(r03): 25(ptr) Variable Function
1173(r04): 37(ptr) Variable Function
1177(r05): 7(ptr) Variable Function
1181(r06): 7(ptr) Variable Function
1185(r07): 37(ptr) Variable Function
1189(r08): 25(ptr) Variable Function
1193(r09): 25(ptr) Variable Function
1197(r10): 37(ptr) Variable Function
1201(r11): 112(ptr) Variable Function
1205(r12): 114(ptr) Variable Function
1209(r13): 61(ptr) Variable Function
1213(r14): 112(ptr) Variable Function
1217(r15): 118(ptr) Variable Function
1221(r16): 116(ptr) Variable Function
1158: 6(float) Load 121(inF1)
1159: 6(float) Load 120(inF0)
1160: 6(float) FMul 1158 1159
Store 1157(r00) 1160
1162: 6(float) Load 120(inF0)
1163: 24(fvec2) Load 122(inFV2)
1164: 24(fvec2) VectorTimesScalar 1163 1162
Store 1161(r01) 1164
1166: 6(float) Load 120(inF0)
1167: 36(fvec3) Load 123(inFV3)
1168: 36(fvec3) VectorTimesScalar 1167 1166
Store 1165(r02) 1168
1170: 24(fvec2) Load 122(inFV2)
1171: 6(float) Load 120(inF0)
1172: 24(fvec2) VectorTimesScalar 1170 1171
Store 1169(r03) 1172
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1174: 36(fvec3) Load 123(inFV3)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1175: 6(float) Load 120(inF0)
1176: 36(fvec3) VectorTimesScalar 1174 1175
Store 1173(r04) 1176
HLSL: add intrinsic function implicit promotions This PR handles implicit promotions for intrinsics when there is no exact match, such as for example clamp(int, bool, float). In this case the int and bool will be promoted to a float, and the clamp(float, float, float) form used. These promotions can be mixed with shape conversions, e.g, clamp(int, bool2, float2). Output conversions are handled either via the existing addOutputArgumentConversion function, which this PR generalizes to handle either aggregates or unaries, or by intrinsic decomposition. If there are methods or intrinsics to be decomposed, then decomposition is responsible for any output conversions, which turns out to happen automatically in all current cases. This can be revisited once inout conversions are in place. Some cases of actual ambiguity were fixed in several tests, e.g, spv.register.autoassign.* Some intrinsics with only uint versions were expanded to signed ints natively, where the underlying AST and SPIR-V supports that. E.g, countbits. This avoids extraneous conversion nodes. A new function promoteAggregate is added, and used by findFunction. This is essentially a generalization of the "promote 1st or 2nd arg" algorithm in promoteBinary. The actual selection proceeds in three steps, as described in the comments in hlslParseContext::findFunction: 1. Attempt an exact match. If found, use it. 2. If not, obtain the operator from step 1, and promote arguments. 3. Re-select the intrinsic overload from the results of step 2.
2016-11-02 18:42:34 +00:00
1178: 24(fvec2) Load 122(inFV2)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1179: 24(fvec2) Load 122(inFV2)
1180: 6(float) Dot 1178 1179
Store 1177(r05) 1180
1182: 36(fvec3) Load 123(inFV3)
1183: 36(fvec3) Load 123(inFV3)
1184: 6(float) Dot 1182 1183
Store 1181(r06) 1184
1186: 111 Load 124(inFM2x3)
1187: 24(fvec2) Load 122(inFV2)
1188: 36(fvec3) MatrixTimesVector 1186 1187
Store 1185(r07) 1188
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1190: 113 Load 125(inFM3x2)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1191: 36(fvec3) Load 123(inFV3)
1192: 24(fvec2) MatrixTimesVector 1190 1191
Store 1189(r08) 1192
1194: 36(fvec3) Load 123(inFV3)
HLSL: Move frexp() to a separate test file.
2017-03-29 09:04:58 +00:00
1195: 111 Load 124(inFM2x3)
HLSL: handle type conversion for any/all intrinsics HLSL allows float/etc types for any/all intrinsics, while the SPIR-V opcode requires bool. This adds a simple decomposition to type convert the argument. It could get a little more clever in some of the type cases if it ever had to.
2017-07-05 18:07:16 +00:00
1196: 24(fvec2) VectorTimesMatrix 1194 1195
Store 1193(r09) 1196
1198: 24(fvec2) Load 122(inFV2)
1199: 113 Load 125(inFM3x2)
1200: 36(fvec3) VectorTimesMatrix 1198 1199
Store 1197(r10) 1200
1202: 6(float) Load 120(inF0)
1203: 111 Load 124(inFM2x3)
1204: 111 MatrixTimesScalar 1203 1202
Store 1201(r11) 1204
1206: 6(float) Load 120(inF0)
1207: 113 Load 125(inFM3x2)
1208: 113 MatrixTimesScalar 1207 1206
Store 1205(r12) 1208
1210: 113 Load 125(inFM3x2)
1211: 111 Load 124(inFM2x3)
1212: 60 MatrixTimesMatrix 1210 1211
Store 1209(r13) 1212
1214: 68 Load 126(inFM3x3)
1215: 111 Load 124(inFM2x3)
1216: 111 MatrixTimesMatrix 1214 1215
Store 1213(r14) 1216
1218: 115 Load 127(inFM3x4)
1219: 111 Load 124(inFM2x3)
1220: 117 MatrixTimesMatrix 1218 1219
Store 1217(r15) 1220
1222: 117 Load 128(inFM2x4)
1223: 113 Load 125(inFM3x2)
1224: 115 MatrixTimesMatrix 1222 1223
Store 1221(r16) 1224
Add decompositions for some HLSL intrinsics.
2016-06-09 14:57:35 +00:00
Return
Initial implementation of direct-mapped subset of HLSL intrinsics with type subset. This checkin implements about half of the HLSL intrinsics for a subset of their entire type support (but a useful subset). The uncommented lines in TBuiltInParseablesHlsl::identifyBuiltIns shows which are connected.
2016-05-26 16:10:16 +00:00
FunctionEnd
Reference in New Issue Copy Permalink