Commit Graph

67 Commits

Author SHA1 Message Date
Hans-Kristian Arntzen
67b2991451 Don't emit memoryBarrierShared() in workgroup control barriers.
This is implied in both GL and GLES. Emitting memoryBarrierShared() was
based on earlier confusion in the spec which has since been fixed and
clarified.
2019-12-04 15:06:19 +01:00
Dan Sinclair
d409210ee5 Move all .invalid shaders into no-opt folders. 2019-11-05 13:19:19 -05:00
Hans-Kristian Arntzen
461f1506e7 Do not eagerly invalidate all active variables on a branch.
This is not necessary, as we must emit an invalidating store before we
potentially consume an invalid expression. In fact, we're a bit
conservative here in this case for example:

int tmp = variable;
if (...)
{
    variable = 10;
}
else
{
    // Consuming tmp here is fine, but it was
    // invalidated while emitting other branch.
    // Technically, we need to study if there is an invalidating store
    // in the CFG between the loading block and this block, and the other
    // branch will not be a part of that analysis.
    int tmp2 = tmp * tmp;
}

Fixing this case means complex CFG traversal *everywhere*, and it feels like overkill.

Fixing this exposed a bug with access chains, so fix a bug where expression dependencies were not
inherited properly in access chains. Access chains are now considered forwarded if there
is at least one dependency which is also forwarded.
2019-07-24 11:17:30 +02:00
Hans-Kristian Arntzen
f8b084de61 MSL/HLSL: Support OpOuterProduct. 2019-07-01 10:57:27 +02:00
Hans-Kristian Arntzen
ff87419607 Deal with scalar input values for distance/length/normalize.
HLSL and MSL don't support it, so fall back to simpler intrinsics.
2019-06-28 11:20:14 +02:00
Hans-Kristian Arntzen
b6f8a20624 GLSL: Return correct sign for OpArrayLength.
.length() returns int, not uint ...
2019-05-07 19:02:32 +02:00
Hans-Kristian Arntzen
e9da5ed631 HLSL: Support OpArrayLength. 2019-05-07 15:53:41 +02:00
Hans-Kristian Arntzen
4ef51331b2 Always value-cast FP16 constants instead of using literals.
GL_NV_gpu_shader5 doesn't support "hf", so to avoid lots of complicated
workarounds, just value-cast the half literals.
2019-02-20 12:30:01 +01:00
Hans-Kristian Arntzen
2ed171e525 GLSL/MSL: Implement 8-bit part of VK_KHR_shader_float16_int8.
Storage was in place already, so mostly just dealing with bitcasts and
constants.

Simplies some of the bitcasting logic, and this exposed some bugs in the
implementation. Refactor to use correct width integers with explicit bitcast opcodes.
2019-01-30 15:45:24 +01:00
Hans-Kristian Arntzen
4e7777c443 Update to latest glslang/SPIRV-Tools.
Fix various bugs along the way.
2019-01-30 13:41:57 +01:00
Hans-Kristian Arntzen
d92de00cc1 Rewrite how IDs are iterated over.
This is a fairly fundamental change on how IDs are handled.
It serves many purposes:

- Improve performance. We only need to iterate over IDs which are
  relevant at any one time.
- Makes sure we iterate through IDs in SPIR-V module declaration order
  rather than ID space. IDs don't have to be monotonically increasing,
  which was an assumption SPIRV-Cross used to have. It has apparently
  never been a problem until now.
- Support LUTs of structs. We do this by interleaving declaration of
  constants and struct types in SPIR-V module order.

To support this, the ParsedIR interface needed to change slightly.
Before setting any ID with variant_set<T> we let ParsedIR know
that an ID with a specific type has been added. The surface for change
should be minimal.

ParsedIR will maintain a per-type list of IDs which the cross-compiler
will need to consider for later.

Instead of looping over ir.ids[] (which can be extremely large), we loop
over types now, using:

ir.for_each_typed_id<SPIRVariable>([&](uint32_t id, SPIRVariable &var) {
	handle_variable(var);
});

Now we make sure that we're never looking at irrelevant types.
2019-01-10 12:52:56 +01:00
Chip Davis
0d949e11ff Support bitcasts of 16-bit types. 2018-11-05 14:56:36 -06:00
Hans-Kristian Arntzen
62db535b3f Update tests. 2018-11-01 11:23:48 +01:00
Grigory Dzhavadyan
a5d82d1138 Alter the handling of spec consts in non-Vulkan GLSL
Previously, when generating non-Vulkan GLSL, each use of a spec constant
would be subsituted for its default value and the declaration of the constant
itself would be omitted completely.

This change slightly alters this behavior. The uses of the constant are kept,
as well as the declaration, although the latter is stripped of the layout
qualifier. The declaration is also prepended with the following code:

    #ifndef <constant name>_value
    #define <constant name> <default constant value>
    #endif

and the constant itself now looks like

    const <constant type> <constant name> = <constant name>_value;

The rationale for this change is that it gives the user a way to provide
custom values for specialization constants even when the target does not
support them.
2018-11-01 00:39:09 -07:00
Hans-Kristian Arntzen
af672b7a4b Add composite array test shaders for GLSL and HLSL as well. 2018-09-10 10:21:08 +02:00
Hans-Kristian Arntzen
d29f48ef06 Deduce constant LUTs from read-write variables. 2018-07-05 13:25:57 +02:00
Hans-Kristian Arntzen
7eba247864 Handle inout properly with split access chains.
Found some other issues. Had some bugs with variable writes not properly
invalidating if writes came from split access chains.
2018-05-11 10:15:42 +02:00
Hans-Kristian Arntzen
85a8f066f4 Do not use RMW rewrite for matrices.
Does not work on MSL.
2018-05-04 10:35:56 +02:00
Hans-Kristian Arntzen
e7bf8d2f48 Refactor out noopt shaders to their own folders.
Makes maintenance easier with less clutter.
2018-03-13 10:39:49 +01:00
Hans-Kristian Arntzen
9fbd8b789e Update tests for latest SPIRV-Tools and glslang. 2018-03-12 15:11:55 +01:00
Hans-Kristian Arntzen
e8e58844d4 Rewrite everything to use Bitset rather than uint64_t. 2018-03-12 13:24:14 +01:00
Hans-Kristian Arntzen
5fe79eb59c Update tests.
Adds an earlier reported shader packing failure into regression suite.
2018-03-05 16:34:42 +01:00
Hans-Kristian Arntzen
27ad8c0922 Add more exhaustive test for barrier handling. 2018-01-09 12:26:46 +01:00
Hans-Kristian Arntzen
166eed8c18 Add test shaders for barriers. 2018-01-09 12:19:31 +01:00
Hans-Kristian Arntzen
7bb8874b43 Optimize away redundant barriers. 2018-01-09 12:17:38 +01:00
Hans-Kristian Arntzen
9c72aa00c9 Overhaul barrier handling in GLSL. 2018-01-09 12:07:07 +01:00
Hans-Kristian Arntzen
23f0abf112 Update tests for struct flattening. 2018-01-09 10:40:34 +01:00
Hans-Kristian Arntzen
3ce6b2a23b Make bitfield test noopt for now.
SPIRV-Tools trips assertion.
2017-11-23 09:59:25 +01:00
Hans-Kristian Arntzen
7e02f7fd62 Check more places where we can potentially read phi variables. 2017-11-23 09:50:11 +01:00
Hans-Kristian Arntzen
6a7b95d2d6 Validate that std140 packing can actually be used. 2017-10-10 10:12:27 +02:00
Hans-Kristian Arntzen
a2229a0327 Add regression test for RMW optimizations. 2017-09-06 09:20:55 +02:00
Hans-Kristian Arntzen
11dfcb6531 Add restrict qualifier for images. 2017-08-29 15:54:22 +02:00
Hans-Kristian Arntzen
d55898ec74 Fix declaration of coherent images. 2017-08-29 15:52:59 +02:00
Hans-Kristian Arntzen
536d888a92 Add test shader for coherent SSBO. 2017-08-28 09:02:08 +02:00
Hans-Kristian Arntzen
744d0405b0 Preserve arguments with inout unless complete writes are made. 2017-08-09 17:06:41 +02:00
Hans-Kristian Arntzen
a4fafa0607 Fix edge-case where do/while body is a dominator.
Shows up quite a lot in inlined code, should get more test coverage in
this area ...
2017-08-02 11:58:24 +02:00
Hans-Kristian Arntzen
b2c2e6483b Analyze parameter preservation for functions.
This is kinda tricky, because if we only conditionally write to a
function parameter variable it is implicitly preserved in SPIR-V, so we must force
an in qualifier on the parameter to get the same behavior in GLSL.
2017-03-25 16:25:30 +01:00
Hans-Kristian Arntzen
24382a8ea2 Use findLSB as well in bitfield test. 2017-03-25 15:46:12 +01:00
Hans-Kristian Arntzen
6801af4dce Fix bitfield functions. 2017-03-25 15:38:20 +01:00
Hans-Kristian Arntzen
0e7c33ff81 Update glslang to latest on Travis.
Fix a lingering bug with OpSelect and boolean mix().
2017-02-11 10:52:34 +01:00
Hans-Kristian Arntzen
d93dc38415 Use correct glslang revision for reference output ... 2017-01-21 12:53:17 +01:00
Hans-Kristian Arntzen
016b1d86e9 Emit readonly, writeonly for SSBOs. 2017-01-21 10:08:27 +01:00
Hans-Kristian Arntzen
d11b8aa3ef Optimize += 1, -= 1 to ++, --.
Purely cosmetic, but easier to read.
2016-12-16 13:24:49 +01:00
Hans-Kristian Arntzen
62613df5a5 Optimize for read-modify-writes.
Required for legacy loop increments.
2016-12-16 13:14:22 +01:00
Hans-Kristian Arntzen
51d45511a6 Check if we can use multiple initializers.
Need same type and qualifiers in GLSL and friends.
2016-12-15 17:54:49 +01:00
Hans-Kristian Arntzen
b8b202f489 Add more dedicated test shader for defer-parens. 2016-12-08 09:05:30 +01:00
Hans-Kristian Arntzen
03a26e593d Fixes for MSL and strip redundant parens in some places. 2016-12-06 23:03:35 +01:00
Hans-Kristian Arntzen
e67f6f85a4 Defer parenthesis generation until needed.
Previously, we would generate parentheses proactively when generating
binary ops, however, this leads to uglier code and hits warnings in
compilers when used as a conditional.
2016-12-05 10:56:54 +01:00
Hans-Kristian Arntzen
2d79d365dc Fix get_declared_struct_size for nested arrays of structs.
The array size was dropped from consideration during recursion.
2016-11-28 15:01:36 +01:00
Hans-Kristian Arntzen
5ff11cc689 Handle more corner cases with the CFG traversal. 2016-11-18 16:45:11 +01:00