Skip to content

v2023.2.0

Compare
Choose a tag to compare
@github-actions github-actions released this 15 Aug 01:29
· 1421 commits to master since this release
b05b126
Slang v2023.2.0

Changes:

b05b126e Update target-deps.json (#3103)
661d6198 Support per field matrix layout (#3101)
0403e055 Fix issue with nested loop unrolling (#3100)
e689d5ee Make sure glsl source in spirv is included when compiling with -g3 (#3099)
37223160 Add scalar version of `distance`. (#3096)
7691128c Upgrade to glslang 12.3.1 (#3095)
756fdb45 Fix `Texture2D.Load()` with offset. (#3094)
38b0af35 Add support for ConstBufferPointer on Vulkan. (#3089)
60ebadab Allow member lookup through modified type. (#3091)
d4e72a93 Support nested structs in function bodies. (#3078)
f875d3f5 Support implciit casted swizzled lvalue. (#3077)
03a5bb4b Various fixes in GLSL emit. (#3074)
c4615fe0 Clean up and improve Val deduplication performance. (#3069)
793a29af Set Full GPU CI timeout to 100 minutes. (#3070)
ea1f1653 Support `-fvk-use-gl-layout` for ByteAddressBuffer load/store. (#3068)
0b2d7533 Add warning on mutation of function parameter (#3067)
97dccb4c Add spirv-dis as a downstream compiler (#3059)
43e0fba1 Default target intrinsic should not apply to SPIR-V (#3054)
ae1c9d98 Fix -fvk-use-entrypoint-name. (#3066)
3a962404 Validate generated SPIR-V on output (#3061)
0e28cd02 Unify Type and Constant sections for SPIR-V (#3057)
03c6cda7 Casting and vector/scalar correct arithmetic ops for SPIR-V (#3056)
0d803a4c Add missing header in slang-glslang (#3055)
10b2038a Diagnostic improvements for direct spirv (#3053)
9eb6a842 Fix `Val` deduplication bug. (#3050)
9ef9cc00 Add missing empty case to Array::makeArray (#3058)
dceb1a74 Allow parsing some SPIR-V enums in intrinsics (#3062)
547ea5cc Simplify SpvTypeInstKey (#3063)
846286bc sv_groupindex support for spirv (#3060)
550396fd Fail on an unhandled spv operand (#3052)
0f57c46a Correct return type for emitFieldAddress (#3051)
856086cf Add option to use original entrypoint in spirv output. (#3047)
a2d90fb2 Redesign `DeclRef` and systematic `Val` deduplication (#3049)
17da4f0d Added D3D12 RenderTargetView TextureCube case (#3048)
77347979 Fix around texture footprint support for NVAPI. (#3046)
aadd816a Disable code motion for expensive insts (call & div) (#3042)
0dcae1c6 Only define atomics for `float2` and `float4` when CUDA arch<900 (#3041)
e80cd4f1 Add mailmap (#3043)
16537317 Fix literals needing cast (#3039)
edcc50cd Update footprint-queries.md
c34a7b66 Generalize collectInductionValues (#3031)
53492410 Fix for #elif evaluation issue (#3038)
8bfdc392 Fix push constant on global variables. (#3034)
04f72443 Fix -fvk-u-shift not applying to global constant buffer. (#3033)
c1216cae Include NonSemanticDebugInfo when calling glslang (#3030)
ba89fc84 Refactor `dmul(This, Differential)` to `dmul<T:Real>(T, Differential)` (#3029)
b8ade05d Fix scalar swizzle causes invalid glsl output.  (#3028)
727245a6 Fix -fvk-u-shift not applying to RWStructuredBuffer on glsl output. (#3027)
7b48b24b Add glsl intrinsics for CalculateLevelOfDetail (#3023)
c61775e6 Add GatherCmp* for texture objects (#3024)
d47d5492 Add slang.natjmc. (#3018)
de3c1b27 Allow parsing more than 10 intrinsic arguments (#3014)
3f64f639 Remove [__readNone] on clip. (#3016)