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
0bb930fc0f
glslang/Test/baseResults/hlsl.samplecmp.offset.dx10.frag.out

586 lines
29 KiB
Plaintext
Raw Normal View History

HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
hlsl.samplecmp.offset.dx10.frag
HLSL: Emit the OpSource HLSL instruction for HLSL, using new headers.
2017-04-07 21:33:08 +00:00
Shader version: 500
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
gl_FragCoord origin is upper left
HLSL/SPV: Declare DepthReplacing for shaders declaring SV_Depth* outputs.
2018-04-14 00:36:58 +00:00
using depth_any
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +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
0:38 Function Definition: @main( ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:38 Function Parameters:
0:? Sequence
0:42 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
0:42 move second child to first child ( temp float)
0:42 'r01' ( temp float)
0:42 textureOffset ( temp float)
0:42 Construct combined texture-sampler ( temp sampler1DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:42 'g_tTex1df4' (layout( binding=0) uniform texture1DShadow)
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 'g_sSamp' (layout( binding=0) uniform sampler)
0:42 Construct vec2 ( temp 2-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:42 Constant:
0:42 0.100000
0:42 Constant:
0:42 0.750000
0:42 Constant:
0:42 2 (const int)
0:43 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
0:43 move second child to first child ( temp float)
0:43 'r03' ( temp float)
0:43 textureOffset ( temp float)
0:43 Construct combined texture-sampler ( temp isampler1DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:43 'g_tTex1di4' ( uniform itexture1DShadow)
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
0:43 'g_sSamp' (layout( binding=0) uniform sampler)
0:43 Construct vec2 ( temp 2-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:43 Constant:
0:43 0.100000
0:43 Constant:
0:43 0.750000
0:43 Constant:
0:43 2 (const int)
0:44 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
0:44 move second child to first child ( temp float)
0:44 'r05' ( temp float)
0:44 textureOffset ( temp float)
0:44 Construct combined texture-sampler ( temp usampler1DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:44 'g_tTex1du4' ( uniform utexture1DShadow)
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
0:44 'g_sSamp' (layout( binding=0) uniform sampler)
0:44 Construct vec2 ( temp 2-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:44 Constant:
0:44 0.100000
0:44 Constant:
0:44 0.750000
0:44 Constant:
0:44 2 (const int)
0:47 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
0:47 move second child to first child ( temp float)
0:47 'r21' ( temp float)
0:47 textureOffset ( temp float)
0:47 Construct combined texture-sampler ( temp sampler2DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:47 'g_tTex2df4' ( uniform texture2DShadow)
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 'g_sSamp' (layout( binding=0) uniform sampler)
0:47 Construct vec3 ( temp 3-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Constant:
0:? 0.100000
0:? 0.200000
0:47 Constant:
0:47 0.750000
0:? Constant:
0:? 2 (const int)
0:? 3 (const int)
0:48 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
0:48 move second child to first child ( temp float)
0:48 'r23' ( temp float)
0:48 textureOffset ( temp float)
0:48 Construct combined texture-sampler ( temp isampler2DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:48 'g_tTex2di4' ( uniform itexture2DShadow)
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 'g_sSamp' (layout( binding=0) uniform sampler)
0:48 Construct vec3 ( temp 3-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Constant:
0:? 0.100000
0:? 0.200000
0:48 Constant:
0:48 0.750000
0:? Constant:
0:? 2 (const int)
0:? 3 (const int)
0:49 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
0:49 move second child to first child ( temp float)
0:49 'r25' ( temp float)
0:49 textureOffset ( temp float)
0:49 Construct combined texture-sampler ( temp usampler2DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:49 'g_tTex2du4' ( uniform utexture2DShadow)
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
0:49 'g_sSamp' (layout( binding=0) uniform sampler)
0:49 Construct vec3 ( temp 3-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Constant:
0:? 0.100000
0:? 0.200000
0:49 Constant:
0:49 0.750000
0:? Constant:
0:? 2 (const int)
0:? 3 (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
0:62 move second child to first child ( temp 4-component vector of float)
0:62 Color: direct index for structure ( temp 4-component vector of float)
0:62 'psout' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Add shape conversions for scalar -> vector assigments. Also, this allows turning on the error check for a failed assigment when parsing. This makes 39 HLSL tests have a working assignment that was previously silently dropped, due to lack of this functionality.
2016-07-28 23:53:56 +00:00
0:62 Constant:
0:62 0 (const int)
0:62 Constant:
0:62 1.000000
0:62 1.000000
0:62 1.000000
0:62 1.000000
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
0:63 move second child to first child ( temp float)
0:63 Depth: direct index for structure ( temp float)
0:63 'psout' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:63 Constant:
0:63 1 (const int)
0:63 Constant:
0:63 1.000000
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:65 Branch: Return with expression
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
0:65 'psout' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 Function Definition: main( ( temp void)
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:38 Function Parameters:
0:? Sequence
0:38 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
0:38 move second child to first child ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 'flattenTemp' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 Function Call: @main( ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 move second child to first child ( temp 4-component vector of float)
HLSL: Move debug naming to a simpler, more consistent, scheme. This will help in expanding flattening and reducing splitting.
2017-08-09 20:29:29 +00:00
0:? '@entryPointOutput.Color' (layout( location=0) out 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
0:38 Color: direct index for structure ( temp 4-component vector of float)
0:38 'flattenTemp' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:38 Constant:
0:38 0 (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
0:38 move second child to first child ( temp float)
HLSL: Move debug naming to a simpler, more consistent, scheme. This will help in expanding flattening and reducing splitting.
2017-08-09 20:29:29 +00:00
0:? '@entryPointOutput.Depth' ( out float FragDepth)
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 Depth: direct index for structure ( temp float)
0:38 'flattenTemp' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:38 Constant:
0:38 1 (const int)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Linker Objects
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:? 'g_sSamp' (layout( binding=0) uniform sampler)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:? 'g_tTex1df4' (layout( binding=0) uniform texture1DShadow)
0:? 'g_tTex1di4' ( uniform itexture1DShadow)
0:? 'g_tTex1du4' ( uniform utexture1DShadow)
0:? 'g_tTex2df4' ( uniform texture2DShadow)
0:? 'g_tTex2di4' ( uniform itexture2DShadow)
0:? 'g_tTex2du4' ( uniform utexture2DShadow)
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:? 'g_tTex3df4' ( uniform texture3D)
0:? 'g_tTex3di4' ( uniform itexture3D)
0:? 'g_tTex3du4' ( uniform utexture3D)
0:? 'g_tTexcdf4' ( uniform textureCube)
0:? 'g_tTexcdi4' ( uniform itextureCube)
0:? 'g_tTexcdu4' ( uniform utextureCube)
0:? 'g_tTex1df4a' ( uniform texture1DArray)
0:? 'g_tTex1di4a' ( uniform itexture1DArray)
0:? 'g_tTex1du4a' ( uniform utexture1DArray)
0:? 'g_tTex2df4a' ( uniform texture2DArray)
0:? 'g_tTex2di4a' ( uniform itexture2DArray)
0:? 'g_tTex2du4a' ( uniform utexture2DArray)
0:? 'g_tTexcdf4a' ( uniform textureCubeArray)
0:? 'g_tTexcdi4a' ( uniform itextureCubeArray)
0:? 'g_tTexcdu4a' ( uniform utextureCubeArray)
HLSL: Move debug naming to a simpler, more consistent, scheme. This will help in expanding flattening and reducing splitting.
2017-08-09 20:29:29 +00:00
0:? '@entryPointOutput.Depth' ( out float FragDepth)
0:? '@entryPointOutput.Color' (layout( location=0) out 4-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
Linked fragment stage:
HLSL: Emit the OpSource HLSL instruction for HLSL, using new headers.
2017-04-07 21:33:08 +00:00
Shader version: 500
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
gl_FragCoord origin is upper left
HLSL/SPV: Declare DepthReplacing for shaders declaring SV_Depth* outputs.
2018-04-14 00:36:58 +00:00
using depth_any
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +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
0:38 Function Definition: @main( ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:38 Function Parameters:
0:? Sequence
0:42 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
0:42 move second child to first child ( temp float)
0:42 'r01' ( temp float)
0:42 textureOffset ( temp float)
0:42 Construct combined texture-sampler ( temp sampler1DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:42 'g_tTex1df4' (layout( binding=0) uniform texture1DShadow)
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 'g_sSamp' (layout( binding=0) uniform sampler)
0:42 Construct vec2 ( temp 2-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:42 Constant:
0:42 0.100000
0:42 Constant:
0:42 0.750000
0:42 Constant:
0:42 2 (const int)
0:43 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
0:43 move second child to first child ( temp float)
0:43 'r03' ( temp float)
0:43 textureOffset ( temp float)
0:43 Construct combined texture-sampler ( temp isampler1DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:43 'g_tTex1di4' ( uniform itexture1DShadow)
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
0:43 'g_sSamp' (layout( binding=0) uniform sampler)
0:43 Construct vec2 ( temp 2-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:43 Constant:
0:43 0.100000
0:43 Constant:
0:43 0.750000
0:43 Constant:
0:43 2 (const int)
0:44 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
0:44 move second child to first child ( temp float)
0:44 'r05' ( temp float)
0:44 textureOffset ( temp float)
0:44 Construct combined texture-sampler ( temp usampler1DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:44 'g_tTex1du4' ( uniform utexture1DShadow)
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
0:44 'g_sSamp' (layout( binding=0) uniform sampler)
0:44 Construct vec2 ( temp 2-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:44 Constant:
0:44 0.100000
0:44 Constant:
0:44 0.750000
0:44 Constant:
0:44 2 (const int)
0:47 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
0:47 move second child to first child ( temp float)
0:47 'r21' ( temp float)
0:47 textureOffset ( temp float)
0:47 Construct combined texture-sampler ( temp sampler2DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:47 'g_tTex2df4' ( uniform texture2DShadow)
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 'g_sSamp' (layout( binding=0) uniform sampler)
0:47 Construct vec3 ( temp 3-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Constant:
0:? 0.100000
0:? 0.200000
0:47 Constant:
0:47 0.750000
0:? Constant:
0:? 2 (const int)
0:? 3 (const int)
0:48 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
0:48 move second child to first child ( temp float)
0:48 'r23' ( temp float)
0:48 textureOffset ( temp float)
0:48 Construct combined texture-sampler ( temp isampler2DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:48 'g_tTex2di4' ( uniform itexture2DShadow)
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 'g_sSamp' (layout( binding=0) uniform sampler)
0:48 Construct vec3 ( temp 3-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Constant:
0:? 0.100000
0:? 0.200000
0:48 Constant:
0:48 0.750000
0:? Constant:
0:? 2 (const int)
0:? 3 (const int)
0:49 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
0:49 move second child to first child ( temp float)
0:49 'r25' ( temp float)
0:49 textureOffset ( temp float)
0:49 Construct combined texture-sampler ( temp usampler2DShadow)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:49 'g_tTex2du4' ( uniform utexture2DShadow)
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
0:49 'g_sSamp' (layout( binding=0) uniform sampler)
0:49 Construct vec3 ( temp 3-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Constant:
0:? 0.100000
0:? 0.200000
0:49 Constant:
0:49 0.750000
0:? Constant:
0:? 2 (const int)
0:? 3 (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
0:62 move second child to first child ( temp 4-component vector of float)
0:62 Color: direct index for structure ( temp 4-component vector of float)
0:62 'psout' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Add shape conversions for scalar -> vector assigments. Also, this allows turning on the error check for a failed assigment when parsing. This makes 39 HLSL tests have a working assignment that was previously silently dropped, due to lack of this functionality.
2016-07-28 23:53:56 +00:00
0:62 Constant:
0:62 0 (const int)
0:62 Constant:
0:62 1.000000
0:62 1.000000
0:62 1.000000
0:62 1.000000
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
0:63 move second child to first child ( temp float)
0:63 Depth: direct index for structure ( temp float)
0:63 'psout' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:63 Constant:
0:63 1 (const int)
0:63 Constant:
0:63 1.000000
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:65 Branch: Return with expression
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
0:65 'psout' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 Function Definition: main( ( temp void)
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:38 Function Parameters:
0:? Sequence
0:38 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
0:38 move second child to first child ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 'flattenTemp' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 Function Call: @main( ( temp structure{ temp 4-component vector of float Color, temp float Depth})
0:38 move second child to first child ( temp 4-component vector of float)
HLSL: Move debug naming to a simpler, more consistent, scheme. This will help in expanding flattening and reducing splitting.
2017-08-09 20:29:29 +00:00
0:? '@entryPointOutput.Color' (layout( location=0) out 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
0:38 Color: direct index for structure ( temp 4-component vector of float)
0:38 'flattenTemp' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:38 Constant:
0:38 0 (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
0:38 move second child to first child ( temp float)
HLSL: Move debug naming to a simpler, more consistent, scheme. This will help in expanding flattening and reducing splitting.
2017-08-09 20:29:29 +00:00
0:? '@entryPointOutput.Depth' ( out float FragDepth)
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 Depth: direct index for structure ( temp float)
0:38 'flattenTemp' ( temp structure{ temp 4-component vector of float Color, temp float Depth})
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
0:38 Constant:
0:38 1 (const int)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
0:? Linker Objects
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:? 'g_sSamp' (layout( binding=0) uniform sampler)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
0:? 'g_tTex1df4' (layout( binding=0) uniform texture1DShadow)
0:? 'g_tTex1di4' ( uniform itexture1DShadow)
0:? 'g_tTex1du4' ( uniform utexture1DShadow)
0:? 'g_tTex2df4' ( uniform texture2DShadow)
0:? 'g_tTex2di4' ( uniform itexture2DShadow)
0:? 'g_tTex2du4' ( uniform utexture2DShadow)
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:? 'g_tTex3df4' ( uniform texture3D)
0:? 'g_tTex3di4' ( uniform itexture3D)
0:? 'g_tTex3du4' ( uniform utexture3D)
0:? 'g_tTexcdf4' ( uniform textureCube)
0:? 'g_tTexcdi4' ( uniform itextureCube)
0:? 'g_tTexcdu4' ( uniform utextureCube)
0:? 'g_tTex1df4a' ( uniform texture1DArray)
0:? 'g_tTex1di4a' ( uniform itexture1DArray)
0:? 'g_tTex1du4a' ( uniform utexture1DArray)
0:? 'g_tTex2df4a' ( uniform texture2DArray)
0:? 'g_tTex2di4a' ( uniform itexture2DArray)
0:? 'g_tTex2du4a' ( uniform utexture2DArray)
0:? 'g_tTexcdf4a' ( uniform textureCubeArray)
0:? 'g_tTexcdi4a' ( uniform itextureCubeArray)
0:? 'g_tTexcdu4a' ( uniform utextureCubeArray)
HLSL: Move debug naming to a simpler, more consistent, scheme. This will help in expanding flattening and reducing splitting.
2017-08-09 20:29:29 +00:00
0:? '@entryPointOutput.Depth' ( out float FragDepth)
0:? '@entryPointOutput.Color' (layout( location=0) out 4-component vector of float)
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
SPV: Add SPIRV-Tools validator. This needs the latest SPIR-Tools.
2018-08-22 23:17:07 +00:00
error: SPIRV-Tools Validation Errors
error: Expected Image 'Sampled Type' to be the same as Result Type
%42 = OpImageSampleDrefImplicitLod %float %39 %40 %41 ConstOffset %int_2
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
Change the major revision number for next commit.
2018-06-04 21:33:31 +00:00
// Generated by (magic number): 80007
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
// Id's are bound by 167
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
Capability Shader
Capability Sampled1D
Capability SampledCubeArray
1: ExtInstImport "GLSL.std.450"
MemoryModel Logical GLSL450
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
EntryPoint Fragment 4 "main" 115 119
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
ExecutionMode 4 OriginUpperLeft
HLSL/SPV: Declare DepthReplacing for shaders declaring SV_Depth* outputs.
2018-04-14 00:36:58 +00:00
ExecutionMode 4 DepthReplacing
HLSL: Emit the OpSource HLSL instruction for HLSL, using new headers.
2017-04-07 21:33:08 +00:00
Source HLSL 500
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
Name 4 "main"
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
Name 8 "PS_OUTPUT"
MemberName 8(PS_OUTPUT) 0 "Color"
MemberName 8(PS_OUTPUT) 1 "Depth"
Name 10 "@main("
Name 13 "r01"
Name 16 "g_tTex1df4"
Name 20 "g_sSamp"
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
Name 32 "r03"
Name 35 "g_tTex1di4"
Name 43 "r05"
Name 47 "g_tTex1du4"
Name 55 "r21"
Name 58 "g_tTex2df4"
Name 74 "r23"
Name 77 "g_tTex2di4"
Name 87 "r25"
Name 90 "g_tTex2du4"
Name 101 "psout"
Name 112 "flattenTemp"
Name 115 "@entryPointOutput.Color"
Name 119 "@entryPointOutput.Depth"
Name 124 "g_tTex3df4"
Name 127 "g_tTex3di4"
Name 130 "g_tTex3du4"
Name 133 "g_tTexcdf4"
Name 136 "g_tTexcdi4"
Name 139 "g_tTexcdu4"
Name 142 "g_tTex1df4a"
Name 145 "g_tTex1di4a"
Name 148 "g_tTex1du4a"
Name 151 "g_tTex2df4a"
Name 154 "g_tTex2di4a"
Name 157 "g_tTex2du4a"
Name 160 "g_tTexcdf4a"
Name 163 "g_tTexcdi4a"
Name 166 "g_tTexcdu4a"
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
Decorate 16(g_tTex1df4) DescriptorSet 0
Decorate 16(g_tTex1df4) Binding 0
Decorate 20(g_sSamp) DescriptorSet 0
Decorate 20(g_sSamp) Binding 0
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
Decorate 35(g_tTex1di4) DescriptorSet 0
Decorate 47(g_tTex1du4) DescriptorSet 0
Decorate 58(g_tTex2df4) DescriptorSet 0
Decorate 77(g_tTex2di4) DescriptorSet 0
Decorate 90(g_tTex2du4) DescriptorSet 0
Decorate 115(@entryPointOutput.Color) Location 0
Decorate 119(@entryPointOutput.Depth) BuiltIn FragDepth
Decorate 124(g_tTex3df4) DescriptorSet 0
Decorate 127(g_tTex3di4) DescriptorSet 0
Decorate 130(g_tTex3du4) DescriptorSet 0
Decorate 133(g_tTexcdf4) DescriptorSet 0
Decorate 136(g_tTexcdi4) DescriptorSet 0
Decorate 139(g_tTexcdu4) DescriptorSet 0
Decorate 142(g_tTex1df4a) DescriptorSet 0
Decorate 145(g_tTex1di4a) DescriptorSet 0
Decorate 148(g_tTex1du4a) DescriptorSet 0
Decorate 151(g_tTex2df4a) DescriptorSet 0
Decorate 154(g_tTex2di4a) DescriptorSet 0
Decorate 157(g_tTex2du4a) DescriptorSet 0
Decorate 160(g_tTexcdf4a) DescriptorSet 0
Decorate 163(g_tTexcdi4a) DescriptorSet 0
Decorate 166(g_tTexcdu4a) DescriptorSet 0
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
2: TypeVoid
3: TypeFunction 2
6: TypeFloat 32
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
7: TypeVector 6(float) 4
8(PS_OUTPUT): TypeStruct 7(fvec4) 6(float)
9: TypeFunction 8(PS_OUTPUT)
12: TypePointer Function 6(float)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
14: TypeImage 6(float) 1D depth sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
15: TypePointer UniformConstant 14
16(g_tTex1df4): 15(ptr) Variable UniformConstant
18: TypeSampler
19: TypePointer UniformConstant 18
20(g_sSamp): 19(ptr) Variable UniformConstant
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
22: TypeSampledImage 14
24: 6(float) Constant 1036831949
25: 6(float) Constant 1061158912
26: TypeVector 6(float) 2
28: TypeInt 32 1
29: 28(int) Constant 2
33: TypeImage 28(int) 1D depth sampled format:Unknown
34: TypePointer UniformConstant 33
35(g_tTex1di4): 34(ptr) Variable UniformConstant
38: TypeSampledImage 33
44: TypeInt 32 0
45: TypeImage 44(int) 1D depth sampled format:Unknown
46: TypePointer UniformConstant 45
47(g_tTex1du4): 46(ptr) Variable UniformConstant
50: TypeSampledImage 45
56: TypeImage 6(float) 2D depth sampled format:Unknown
57: TypePointer UniformConstant 56
58(g_tTex2df4): 57(ptr) Variable UniformConstant
61: TypeSampledImage 56
63: 6(float) Constant 1045220557
64: 26(fvec2) ConstantComposite 24 63
65: TypeVector 6(float) 3
69: TypeVector 28(int) 2
70: 28(int) Constant 3
71: 69(ivec2) ConstantComposite 29 70
75: TypeImage 28(int) 2D depth sampled format:Unknown
76: TypePointer UniformConstant 75
77(g_tTex2di4): 76(ptr) Variable UniformConstant
80: TypeSampledImage 75
88: TypeImage 44(int) 2D depth sampled format:Unknown
89: TypePointer UniformConstant 88
90(g_tTex2du4): 89(ptr) Variable UniformConstant
93: TypeSampledImage 88
100: TypePointer Function 8(PS_OUTPUT)
102: 28(int) Constant 0
103: 6(float) Constant 1065353216
104: 7(fvec4) ConstantComposite 103 103 103 103
105: TypePointer Function 7(fvec4)
107: 28(int) Constant 1
114: TypePointer Output 7(fvec4)
115(@entryPointOutput.Color): 114(ptr) Variable Output
118: TypePointer Output 6(float)
119(@entryPointOutput.Depth): 118(ptr) Variable Output
122: TypeImage 6(float) 3D sampled format:Unknown
123: TypePointer UniformConstant 122
124(g_tTex3df4): 123(ptr) Variable UniformConstant
125: TypeImage 28(int) 3D sampled format:Unknown
126: TypePointer UniformConstant 125
127(g_tTex3di4): 126(ptr) Variable UniformConstant
128: TypeImage 44(int) 3D sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
129: TypePointer UniformConstant 128
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
130(g_tTex3du4): 129(ptr) Variable UniformConstant
131: TypeImage 6(float) Cube sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
132: TypePointer UniformConstant 131
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
133(g_tTexcdf4): 132(ptr) Variable UniformConstant
134: TypeImage 28(int) Cube sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
135: TypePointer UniformConstant 134
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
136(g_tTexcdi4): 135(ptr) Variable UniformConstant
137: TypeImage 44(int) Cube sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
138: TypePointer UniformConstant 137
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
139(g_tTexcdu4): 138(ptr) Variable UniformConstant
140: TypeImage 6(float) 1D array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
141: TypePointer UniformConstant 140
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
142(g_tTex1df4a): 141(ptr) Variable UniformConstant
143: TypeImage 28(int) 1D array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
144: TypePointer UniformConstant 143
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
145(g_tTex1di4a): 144(ptr) Variable UniformConstant
146: TypeImage 44(int) 1D array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
147: TypePointer UniformConstant 146
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
148(g_tTex1du4a): 147(ptr) Variable UniformConstant
149: TypeImage 6(float) 2D array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
150: TypePointer UniformConstant 149
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
151(g_tTex2df4a): 150(ptr) Variable UniformConstant
152: TypeImage 28(int) 2D array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
153: TypePointer UniformConstant 152
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
154(g_tTex2di4a): 153(ptr) Variable UniformConstant
155: TypeImage 44(int) 2D array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
156: TypePointer UniformConstant 155
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
157(g_tTex2du4a): 156(ptr) Variable UniformConstant
158: TypeImage 6(float) Cube array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
159: TypePointer UniformConstant 158
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
160(g_tTexcdf4a): 159(ptr) Variable UniformConstant
161: TypeImage 28(int) Cube array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
162: TypePointer UniformConstant 161
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
163(g_tTexcdi4a): 162(ptr) Variable UniformConstant
164: TypeImage 44(int) Cube array sampled format:Unknown
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
165: TypePointer UniformConstant 164
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
166(g_tTexcdu4a): 165(ptr) Variable UniformConstant
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
4(main): 2 Function None 3
5: Label
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
112(flattenTemp): 100(ptr) Variable Function
113:8(PS_OUTPUT) FunctionCall 10(@main()
Store 112(flattenTemp) 113
116: 105(ptr) AccessChain 112(flattenTemp) 102
117: 7(fvec4) Load 116
Store 115(@entryPointOutput.Color) 117
120: 12(ptr) AccessChain 112(flattenTemp) 107
121: 6(float) Load 120
Store 119(@entryPointOutput.Depth) 121
HLSL: Emulate write-to-output on return-from-entry-point, for return value. This fixes issue #487 and #480. It also correctly handles output parameters from the entry point.
2016-08-28 21:00:23 +00:00
Return
HLSL: Add SampleCmp and SampleCmpLevelZero texture methods
2016-07-19 20:28:05 +00:00
FunctionEnd
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
10(@main():8(PS_OUTPUT) Function None 9
11: Label
13(r01): 12(ptr) Variable Function
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
32(r03): 12(ptr) Variable Function
43(r05): 12(ptr) Variable Function
55(r21): 12(ptr) Variable Function
74(r23): 12(ptr) Variable Function
87(r25): 12(ptr) Variable Function
101(psout): 100(ptr) Variable Function
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
17: 14 Load 16(g_tTex1df4)
21: 18 Load 20(g_sSamp)
HLSL: force textures to shadow modes from combined samplers Texture shadow mode must match the state of the sampler they are combined with. This change does that, both for the AST and the symbol table. Note that the texture cannot easily be *created* the right way, because this may not be known at that time. Instead, the texture is subsequently patched. This cannot work if a single texture is used with both a shadow and non-shadow sampler, so that case is detected and generates an error. This is permitted by the HLSL language, however. See #1073 discussion. Fixed one test source that was using a texture with both shadow and non-shadow samplers.
2017-09-27 15:12:51 +00:00
23: 22 SampledImage 17 21
27: 26(fvec2) CompositeConstruct 24 25
30: 6(float) CompositeExtract 27 1
31: 6(float) ImageSampleDrefImplicitLod 23 27 30 ConstOffset 29
Store 13(r01) 31
36: 33 Load 35(g_tTex1di4)
37: 18 Load 20(g_sSamp)
39: 38 SampledImage 36 37
40: 26(fvec2) CompositeConstruct 24 25
41: 6(float) CompositeExtract 40 1
42: 6(float) ImageSampleDrefImplicitLod 39 40 41 ConstOffset 29
Store 32(r03) 42
48: 45 Load 47(g_tTex1du4)
49: 18 Load 20(g_sSamp)
51: 50 SampledImage 48 49
52: 26(fvec2) CompositeConstruct 24 25
53: 6(float) CompositeExtract 52 1
54: 6(float) ImageSampleDrefImplicitLod 51 52 53 ConstOffset 29
Store 43(r05) 54
59: 56 Load 58(g_tTex2df4)
60: 18 Load 20(g_sSamp)
62: 61 SampledImage 59 60
66: 6(float) CompositeExtract 64 0
67: 6(float) CompositeExtract 64 1
68: 65(fvec3) CompositeConstruct 66 67 25
72: 6(float) CompositeExtract 68 2
73: 6(float) ImageSampleDrefImplicitLod 62 68 72 ConstOffset 71
Store 55(r21) 73
78: 75 Load 77(g_tTex2di4)
79: 18 Load 20(g_sSamp)
81: 80 SampledImage 78 79
82: 6(float) CompositeExtract 64 0
83: 6(float) CompositeExtract 64 1
84: 65(fvec3) CompositeConstruct 82 83 25
85: 6(float) CompositeExtract 84 2
86: 6(float) ImageSampleDrefImplicitLod 81 84 85 ConstOffset 71
Store 74(r23) 86
91: 88 Load 90(g_tTex2du4)
92: 18 Load 20(g_sSamp)
94: 93 SampledImage 91 92
95: 6(float) CompositeExtract 64 0
96: 6(float) CompositeExtract 64 1
97: 65(fvec3) CompositeConstruct 95 96 25
98: 6(float) CompositeExtract 97 2
99: 6(float) ImageSampleDrefImplicitLod 94 97 98 ConstOffset 71
Store 87(r25) 99
106: 105(ptr) AccessChain 101(psout) 102
Store 106 104
108: 12(ptr) AccessChain 101(psout) 107
Store 108 103
109:8(PS_OUTPUT) Load 101(psout)
ReturnValue 109
HLSL: Wrap the entry-point; need to write 'in' args, and support 'inout' args. This needs some render testing, but is destined to be part of master. This also leads to a variety of other simplifications. - IO are global symbols, so only need one list of linkage nodes (deferred) - no longer need parse-context-wide 'inEntryPoint' state, entry-point is localized - several parts of splitting/flattening are now localized
2017-01-19 22:41:47 +00:00
FunctionEnd
Reference in New Issue Copy Permalink