{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":136128198,"defaultBranch":"dev","name":"llpc","ownerLogin":"GPUOpen-Drivers","currentUserCanPush":false,"isFork":false,"isEmpty":false,"createdAt":"2018-06-05T06:06:02.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/29980581?v=4","public":true,"private":false,"isOrgOwned":true},"refInfo":{"name":"","listCacheKey":"v0:1706249266.0","currentOid":""},"activityList":{"items":[{"before":"6c770c7d276d2c2504aed2a0278aab1610993ecf","after":"f64d106c023e7bed59bfb441ffb258c127e500c1","ref":"refs/heads/master","pushedAt":"2024-06-26T08:25:56.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 879e8809\n\n[CONTINUATIONS] Add metadata required by RGP\n[CONTINUATIONS] Run some optimization pass for gpurt module\nAdd NoContraction decoration to have a test for fmul_legacy mapping\nAdd readfirstlane on the result of subgroupClusterReduction on gfx11+\nAdd RobustGsEmits to GFX10\nAdd support for GS patch primitive type\namdllpc: Add more helpful info with PipelineLib* dumps\nAvoid upgrade to seqcst ordering\nDebug Printf refactor\nDowngrade SequentiallyConsistent to AcquireRelease\nExpose getResourceMappingNodeTypeName() to the driver\nlgc: Improve TANH expansion to avoid overflow\nlgc: Refactor getShaderStageAbbreviation\nlgc: Use agent scope in more places\nllpcSpirvLowerGlobal: Fix originUpperLeft handling\nllpcSpirvLowerGlobal: Refactor input/output lowering\nllvmraytracing: Separate out header file for pointee type metadata\nPostpone descriptor load to ImageBuilder\nPromote llvm-dialects submodule\nSet last-use for load from the continuation stack\nSupport for LogRayTracingPipelineSummary\nUpdate shader tests after LLVM update\nFix a typo on paClVsOutCntl\nFix primitive type for barycentric\nFix the cooperativematrix issues(convert+muladd) on gfx1010\nFixes for lit tests on standalone amdllpc build","shortMessageHtmlLink":"Update llpc from commit 879e8809"}},{"before":"6c770c7d276d2c2504aed2a0278aab1610993ecf","after":"f64d106c023e7bed59bfb441ffb258c127e500c1","ref":"refs/heads/dev","pushedAt":"2024-06-26T08:14:17.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 879e8809\n\n[CONTINUATIONS] Add metadata required by RGP\n[CONTINUATIONS] Run some optimization pass for gpurt module\nAdd NoContraction decoration to have a test for fmul_legacy mapping\nAdd readfirstlane on the result of subgroupClusterReduction on gfx11+\nAdd RobustGsEmits to GFX10\nAdd support for GS patch primitive type\namdllpc: Add more helpful info with PipelineLib* dumps\nAvoid upgrade to seqcst ordering\nDebug Printf refactor\nDowngrade SequentiallyConsistent to AcquireRelease\nExpose getResourceMappingNodeTypeName() to the driver\nlgc: Improve TANH expansion to avoid overflow\nlgc: Refactor getShaderStageAbbreviation\nlgc: Use agent scope in more places\nllpcSpirvLowerGlobal: Fix originUpperLeft handling\nllpcSpirvLowerGlobal: Refactor input/output lowering\nllvmraytracing: Separate out header file for pointee type metadata\nPostpone descriptor load to ImageBuilder\nPromote llvm-dialects submodule\nSet last-use for load from the continuation stack\nSupport for LogRayTracingPipelineSummary\nUpdate shader tests after LLVM update\nFix a typo on paClVsOutCntl\nFix primitive type for barycentric\nFix the cooperativematrix issues(convert+muladd) on gfx1010\nFixes for lit tests on standalone amdllpc build","shortMessageHtmlLink":"Update llpc from commit 879e8809"}},{"before":"6c770c7d276d2c2504aed2a0278aab1610993ecf","after":"f64d106c023e7bed59bfb441ffb258c127e500c1","ref":"refs/heads/dev_promote","pushedAt":"2024-06-26T06:02:00.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 879e8809\n\n[CONTINUATIONS] Add metadata required by RGP\n[CONTINUATIONS] Run some optimization pass for gpurt module\nAdd NoContraction decoration to have a test for fmul_legacy mapping\nAdd readfirstlane on the result of subgroupClusterReduction on gfx11+\nAdd RobustGsEmits to GFX10\nAdd support for GS patch primitive type\namdllpc: Add more helpful info with PipelineLib* dumps\nAvoid upgrade to seqcst ordering\nDebug Printf refactor\nDowngrade SequentiallyConsistent to AcquireRelease\nExpose getResourceMappingNodeTypeName() to the driver\nlgc: Improve TANH expansion to avoid overflow\nlgc: Refactor getShaderStageAbbreviation\nlgc: Use agent scope in more places\nllpcSpirvLowerGlobal: Fix originUpperLeft handling\nllpcSpirvLowerGlobal: Refactor input/output lowering\nllvmraytracing: Separate out header file for pointee type metadata\nPostpone descriptor load to ImageBuilder\nPromote llvm-dialects submodule\nSet last-use for load from the continuation stack\nSupport for LogRayTracingPipelineSummary\nUpdate shader tests after LLVM update\nFix a typo on paClVsOutCntl\nFix primitive type for barycentric\nFix the cooperativematrix issues(convert+muladd) on gfx1010\nFixes for lit tests on standalone amdllpc build","shortMessageHtmlLink":"Update llpc from commit 879e8809"}},{"before":"8252cfa83e03aa0a193c0f0cd6a9dba3e99e25d5","after":"6c770c7d276d2c2504aed2a0278aab1610993ecf","ref":"refs/heads/master","pushedAt":"2024-06-05T03:31:41.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit e48a029e\n\n[Continuations] Add a test with boolean payload\n[Continuations] Add assertion on payload register count usage\n[Continuations] Add getUpperBoundOnTraceRayPayloadRegisters()\n[Continuations] Add shader jump scheduling workflow\n[Continuations] Add simple lit test for `_cont_ExitRayGen` inlining\n[Continuations] Add stable returnaddr argument\n[Continuations] Cleanup return in `else` block\n[Continuations] Cleanup use of `EntriesWithContinuationsFunctions`\n[Continuations] Dialectify continuation.return.\n[Continuations] Fix a regression caused by strip_whitelist.txt clean-up in GPURT\n[Continuations] Fix Cps Stack Lowering for KernelEntry\n[Continuations] Fix lit test for old LLVM\n[Continuations] Fix Traversal register count metadata\n[Continuations] Introduce `llvm::replaceCallsToFunction` helper\n[Continuations] Introduce `PayloadHelper`\n[Continuations] Make dialect stackstore take no pointer\n[Continuations] Move payload intrinsic handling to `LowerRaytracingPipeline`\n[Continuations] Move static helper function\n[Continuations] PAQ: Represent uniform structs by a single PAQ node\n[Continuations] Pass csp argument to Raygen, remove `_cont_SetupRayGen`\n[Continuations] Refactor `LowerAwait` pass\n[Continuations] Refactor PAQ node creation\n[Continuations] Remove unused includes\n[Continuations] Replace `forEachCall` with call to `replaceCallsToFunction`\n[Continuations] Return `PreservedAnalyes` directly.\n[Continuations] Run `LowerRaytracingPipeline` on Traversal\n[Continuations] Set preserved register count for IS\n[Continuations] Support _cont_ShaderStart\n[Continuations] Support GPURT options\n[Continuations] Use --lint-abort-on-error in lit tests\n[Continuations] Use as.continuation.reference consistently\n[Continuations] Use lgc.rt max payload size metadata\n[Continuations] Use max hit attributes from module\n[Continuations] Use updated update_test_checks.py to generate lit files\nAdd SqBufRsrcWord3 initialization\nAdd support for AmdExtD3DShaderIntrinsics_WaveScan Intrinsic\nEnable scalarization of descriptor loads\nFix AmdTraceRayGetStaticId in ray query\nFix invalid startswith call\nFix static check error: Using 'memset' on struct that contains a 'std::string'\nFix the issue where a maxVertOut value of zero leads to a division by maxVertOut, causing a crash\nHandle upstream llvm changes for metadata\nlgc: handle indirect load case in stridedIndexAdd\nlgc: Add BufferAddrToPtrOp and ExtendAddressOp\nlgc: Add BuilderCommon::CreateBuildVector\nlgc: Add debugging log for PatchBufferOp\nlgc: Add offset and ascii chars comment to bytes of data line\nlgc: emit amdgcn.fdot2.f16.f16 intrinsics\nlgc: Enable sinking in SimplifyCFG\nlgc: Extend lgc dialect with subgroupAll, AllEqual, Rotate\nlgc: Fix 'Add DbgInfo api to ModuleBunch'\nlgc: Fix crash in HLK Dot2AddHalfTest on GFX10.1\nlgc: Fix IO error if pre-rast shader don't export layer\nlgc: Fix the hang when shaderDBG is ON\nlgc: Move CreateIf() to BuilderCommon so a front-end can use it\nlgc: Refactor CreateSubgroupClustered* with contant cluster size\nlgc: Some coding style changes in PatchEntryPointMutate\nlgc: Use rtz version 16bit interpolation\nllvmraytracing: Allow setLgcRtShaderStage/getLgcRtShaderStage on GlobalObject\nRefactor for CooperativeMatrix using LGC_Dialect\nRemove the macro VKI_COOPERATIVE_MATRIX gating under lgc/llpc test cases\nReplace struct_buffer_load with s_buffer_load for strided buffer in some case\nRevert 'Pack mesh shader outputs in LDS space' to fix corruption in The Talos Principle 2\nRevert 'lgc: Add LoadBufferAddrOp'\nRework SW stream-out\nSome coding changes for dynamic topology\nSPIRVReader: Map spirv device scope to agent\nSupport dumping and reading GPURT options\nSupport dynamic primitive topology when xfb is enabled\nSupport uint64 AmdTraceRaySampleGpuTimer()\nUpdate tests affected by change for range attr in ctpop\nUpdate tests for upstream GEP change","shortMessageHtmlLink":"Update llpc from commit e48a029e"}},{"before":"8252cfa83e03aa0a193c0f0cd6a9dba3e99e25d5","after":"6c770c7d276d2c2504aed2a0278aab1610993ecf","ref":"refs/heads/dev","pushedAt":"2024-06-04T07:33:51.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit e48a029e\n\n[Continuations] Add a test with boolean payload\n[Continuations] Add assertion on payload register count usage\n[Continuations] Add getUpperBoundOnTraceRayPayloadRegisters()\n[Continuations] Add shader jump scheduling workflow\n[Continuations] Add simple lit test for `_cont_ExitRayGen` inlining\n[Continuations] Add stable returnaddr argument\n[Continuations] Cleanup return in `else` block\n[Continuations] Cleanup use of `EntriesWithContinuationsFunctions`\n[Continuations] Dialectify continuation.return.\n[Continuations] Fix a regression caused by strip_whitelist.txt clean-up in GPURT\n[Continuations] Fix Cps Stack Lowering for KernelEntry\n[Continuations] Fix lit test for old LLVM\n[Continuations] Fix Traversal register count metadata\n[Continuations] Introduce `llvm::replaceCallsToFunction` helper\n[Continuations] Introduce `PayloadHelper`\n[Continuations] Make dialect stackstore take no pointer\n[Continuations] Move payload intrinsic handling to `LowerRaytracingPipeline`\n[Continuations] Move static helper function\n[Continuations] PAQ: Represent uniform structs by a single PAQ node\n[Continuations] Pass csp argument to Raygen, remove `_cont_SetupRayGen`\n[Continuations] Refactor `LowerAwait` pass\n[Continuations] Refactor PAQ node creation\n[Continuations] Remove unused includes\n[Continuations] Replace `forEachCall` with call to `replaceCallsToFunction`\n[Continuations] Return `PreservedAnalyes` directly.\n[Continuations] Run `LowerRaytracingPipeline` on Traversal\n[Continuations] Set preserved register count for IS\n[Continuations] Support _cont_ShaderStart\n[Continuations] Support GPURT options\n[Continuations] Use --lint-abort-on-error in lit tests\n[Continuations] Use as.continuation.reference consistently\n[Continuations] Use lgc.rt max payload size metadata\n[Continuations] Use max hit attributes from module\n[Continuations] Use updated update_test_checks.py to generate lit files\nAdd SqBufRsrcWord3 initialization\nAdd support for AmdExtD3DShaderIntrinsics_WaveScan Intrinsic\nEnable scalarization of descriptor loads\nFix AmdTraceRayGetStaticId in ray query\nFix invalid startswith call\nFix static check error: Using 'memset' on struct that contains a 'std::string'\nFix the issue where a maxVertOut value of zero leads to a division by maxVertOut, causing a crash\nHandle upstream llvm changes for metadata\nlgc: handle indirect load case in stridedIndexAdd\nlgc: Add BufferAddrToPtrOp and ExtendAddressOp\nlgc: Add BuilderCommon::CreateBuildVector\nlgc: Add debugging log for PatchBufferOp\nlgc: Add offset and ascii chars comment to bytes of data line\nlgc: emit amdgcn.fdot2.f16.f16 intrinsics\nlgc: Enable sinking in SimplifyCFG\nlgc: Extend lgc dialect with subgroupAll, AllEqual, Rotate\nlgc: Fix 'Add DbgInfo api to ModuleBunch'\nlgc: Fix crash in HLK Dot2AddHalfTest on GFX10.1\nlgc: Fix IO error if pre-rast shader don't export layer\nlgc: Fix the hang when shaderDBG is ON\nlgc: Move CreateIf() to BuilderCommon so a front-end can use it\nlgc: Refactor CreateSubgroupClustered* with contant cluster size\nlgc: Some coding style changes in PatchEntryPointMutate\nlgc: Use rtz version 16bit interpolation\nllvmraytracing: Allow setLgcRtShaderStage/getLgcRtShaderStage on GlobalObject\nRefactor for CooperativeMatrix using LGC_Dialect\nRemove the macro VKI_COOPERATIVE_MATRIX gating under lgc/llpc test cases\nReplace struct_buffer_load with s_buffer_load for strided buffer in some case\nRevert 'Pack mesh shader outputs in LDS space' to fix corruption in The Talos Principle 2\nRevert 'lgc: Add LoadBufferAddrOp'\nRework SW stream-out\nSome coding changes for dynamic topology\nSPIRVReader: Map spirv device scope to agent\nSupport dumping and reading GPURT options\nSupport dynamic primitive topology when xfb is enabled\nSupport uint64 AmdTraceRaySampleGpuTimer()\nUpdate tests affected by change for range attr in ctpop\nUpdate tests for upstream GEP change","shortMessageHtmlLink":"Update llpc from commit e48a029e"}},{"before":"538debaf2be990e029d5b8e1ebf3a330af9fc80c","after":"6c770c7d276d2c2504aed2a0278aab1610993ecf","ref":"refs/heads/dev_promote","pushedAt":"2024-06-04T03:49:06.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit e48a029e\n\n[Continuations] Add a test with boolean payload\n[Continuations] Add assertion on payload register count usage\n[Continuations] Add getUpperBoundOnTraceRayPayloadRegisters()\n[Continuations] Add shader jump scheduling workflow\n[Continuations] Add simple lit test for `_cont_ExitRayGen` inlining\n[Continuations] Add stable returnaddr argument\n[Continuations] Cleanup return in `else` block\n[Continuations] Cleanup use of `EntriesWithContinuationsFunctions`\n[Continuations] Dialectify continuation.return.\n[Continuations] Fix a regression caused by strip_whitelist.txt clean-up in GPURT\n[Continuations] Fix Cps Stack Lowering for KernelEntry\n[Continuations] Fix lit test for old LLVM\n[Continuations] Fix Traversal register count metadata\n[Continuations] Introduce `llvm::replaceCallsToFunction` helper\n[Continuations] Introduce `PayloadHelper`\n[Continuations] Make dialect stackstore take no pointer\n[Continuations] Move payload intrinsic handling to `LowerRaytracingPipeline`\n[Continuations] Move static helper function\n[Continuations] PAQ: Represent uniform structs by a single PAQ node\n[Continuations] Pass csp argument to Raygen, remove `_cont_SetupRayGen`\n[Continuations] Refactor `LowerAwait` pass\n[Continuations] Refactor PAQ node creation\n[Continuations] Remove unused includes\n[Continuations] Replace `forEachCall` with call to `replaceCallsToFunction`\n[Continuations] Return `PreservedAnalyes` directly.\n[Continuations] Run `LowerRaytracingPipeline` on Traversal\n[Continuations] Set preserved register count for IS\n[Continuations] Support _cont_ShaderStart\n[Continuations] Support GPURT options\n[Continuations] Use --lint-abort-on-error in lit tests\n[Continuations] Use as.continuation.reference consistently\n[Continuations] Use lgc.rt max payload size metadata\n[Continuations] Use max hit attributes from module\n[Continuations] Use updated update_test_checks.py to generate lit files\nAdd SqBufRsrcWord3 initialization\nAdd support for AmdExtD3DShaderIntrinsics_WaveScan Intrinsic\nEnable scalarization of descriptor loads\nFix AmdTraceRayGetStaticId in ray query\nFix invalid startswith call\nFix static check error: Using 'memset' on struct that contains a 'std::string'\nFix the issue where a maxVertOut value of zero leads to a division by maxVertOut, causing a crash\nHandle upstream llvm changes for metadata\nlgc: handle indirect load case in stridedIndexAdd\nlgc: Add BufferAddrToPtrOp and ExtendAddressOp\nlgc: Add BuilderCommon::CreateBuildVector\nlgc: Add debugging log for PatchBufferOp\nlgc: Add offset and ascii chars comment to bytes of data line\nlgc: emit amdgcn.fdot2.f16.f16 intrinsics\nlgc: Enable sinking in SimplifyCFG\nlgc: Extend lgc dialect with subgroupAll, AllEqual, Rotate\nlgc: Fix 'Add DbgInfo api to ModuleBunch'\nlgc: Fix crash in HLK Dot2AddHalfTest on GFX10.1\nlgc: Fix IO error if pre-rast shader don't export layer\nlgc: Fix the hang when shaderDBG is ON\nlgc: Move CreateIf() to BuilderCommon so a front-end can use it\nlgc: Refactor CreateSubgroupClustered* with contant cluster size\nlgc: Some coding style changes in PatchEntryPointMutate\nlgc: Use rtz version 16bit interpolation\nllvmraytracing: Allow setLgcRtShaderStage/getLgcRtShaderStage on GlobalObject\nRefactor for CooperativeMatrix using LGC_Dialect\nRemove the macro VKI_COOPERATIVE_MATRIX gating under lgc/llpc test cases\nReplace struct_buffer_load with s_buffer_load for strided buffer in some case\nRevert 'Pack mesh shader outputs in LDS space' to fix corruption in The Talos Principle 2\nRevert 'lgc: Add LoadBufferAddrOp'\nRework SW stream-out\nSome coding changes for dynamic topology\nSPIRVReader: Map spirv device scope to agent\nSupport dumping and reading GPURT options\nSupport dynamic primitive topology when xfb is enabled\nSupport uint64 AmdTraceRaySampleGpuTimer()\nUpdate tests affected by change for range attr in ctpop\nUpdate tests for upstream GEP change","shortMessageHtmlLink":"Update llpc from commit e48a029e"}},{"before":"0e7e05f513204d22d125d439560541aaeb664403","after":"538debaf2be990e029d5b8e1ebf3a330af9fc80c","ref":"refs/heads/dev_promote","pushedAt":"2024-05-20T03:59:03.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix the corruption in The Talos Principle2.","shortMessageHtmlLink":"Fix the corruption in The Talos Principle2."}},{"before":"8252cfa83e03aa0a193c0f0cd6a9dba3e99e25d5","after":"0e7e05f513204d22d125d439560541aaeb664403","ref":"refs/heads/dev_promote","pushedAt":"2024-05-20T03:39:24.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix the corruption in The Talos Principle2.","shortMessageHtmlLink":"Fix the corruption in The Talos Principle2."}},{"before":"3f33e94c6faf594537023ccbdafc0ca45dc0315d","after":"8252cfa83e03aa0a193c0f0cd6a9dba3e99e25d5","ref":"refs/heads/dev_promote","pushedAt":"2024-05-20T03:25:49.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 39ee3a1e\n\n[Continuations] Add helper commonMaterializable()\n[Continuations] Add one extra SROA\n[Continuations] Add PostProcessing test.\n[Continuations] Check that resume functions report no stack usage\n[Continuations] Cleanup includes.\n[Continuations] Fix corruption seen in RaytracingReflections\n[Continuations] Fix first payload argument dword calculation\n[Continuations] Fix outgoing register count for kernelentry\n[Continuations] Get DispatchSystemData type from GetLocalRootIndex.\n[Continuations] Handle gpurt GetSetting\n[Continuations] Ignore life time markers when doing store-to-load forwarding\n[Continuations] Keep _AmdAcceptHitAttributes calls\n[Continuations] Lower _GetResumePoint correctly\n[Continuations] Move stackptr lowering\n[Continuations] Promote _AmdAwait arguments correctly\n[Continuations] Refactor RT shader stage enum\n[Continuations] Rematerialize single source shuffle\n[Continuations] Remove old unused helper\n[Continuations] Remove use of [[no_unique_address]]\n[Continuations] Set up proper register count for Traversal\n[Continuations] Spill in PayloadCopyHelper\n[Continuations] Support TriangleVertexPositions in continuations\n[Continuations] Update test to reduce diff\n[Continuations] Use int as name for gpurt settings\n[Continuations] Use llvmraytracing in more places\n[RT] Set the entry name as modulename\nAdd 'override' to SPIRVBaseVariable::validate()\nAdd AmdExtD3DShaderIntrinsics_ShaderMarker\nAdd DbgInfo api to ModuleBunch\nAdd detecting more patterns which can be lowered into fmul_legacy and fma_legacy\nAdd dynamicVgprBlockSize option\nAdd GpurtVersionFlags\nAdd Phoenix2 support\nAdd ShaderModuleUsage::usePrimitiveId\nApply WQM to result of WWM subgroup operations\nClean GEP formation for NGG atomic add\nCleanup the old PAL metadata code path\nCrossModuleInliner: add check for consistent target module\nCrossModuleInliner: make struct layout independent of NDEBUG\nDataLayout index sizes for SPIRV address spaces\nDetect read-only buffers and mark them as invariant\nExpand GEP compatibility in tests\nHandle new return type for some IRBuilder methods\nHandle new return type from DIBuilder\nlgc: add strided buffer load\nlgc: correct stage detection for subgroup builder\nlgc: missing code for load.strided.buffer.desc\nlgc: Add LoadBufferAddrOp\nlgc: Allow hoisting common instructions\nlgc: Disassemble archive-of-ELFs\nlgc: move CreateDebugBreak to BuilderCommon\nlgc: Set flat workgroup size for cs_chain\nLowerRayTracingPipeline: Avoid redundant GEPs\nMove replaceAllPointerUses into CompilerUtils\nPack mesh shader outputs in LDS space\nReduce SW XFB LDS usage\nRefine advanced blend func with value-type arguments\nRefine dump info for shader module id\nRefine pipeline dump when shader module identifier is used\nRemove checks for gfxip >= 10\nRemove LDS granularity rounding\nRemove ShaderModuleUsage::useGenericBuiltIn\nRemove the accidentally-added file for *.pipe.swp\nReport new RequireFullPipeline error\nReturn frozen poison value in _AmdGetUninitialized\nReturn the state whether the texture is accessed by texelFetch function\nSet MaxOutgoingVgprCount into SGPR and pass it\nSimplify LDS GEP formation\nSome renaming in mesh shader\nSupport uint64 AmdTraceRaySampleGpuTimer()\nTargetFeatures for amdgpu_cs_chain_preserve funcs\nUpdate llvm-dialects submodule\nUse 32-bit GEPs consistently\nUse CmpInst::Create instead of new to create instruction\nUse convertUsersOfConstantsToInstruction throughout\nVertexFetch: vertex offset is added twice in vertex offset mode\nFix check_llpc test regression\nFix color export offline compilation\nFix compiler warnings\nFix error on unused private field\nFix ShaderDbg issues\nFix the check for reserving a dummy vertex attribute\nFix unusual case imported by new game:STARFIELD\nFix updating threshold issue when merge with pipeline cache which cause PAL assert","shortMessageHtmlLink":"Update llpc from commit 39ee3a1e"}},{"before":"14535c028dc70a093d95a95efedb749812c1e77d","after":"8252cfa83e03aa0a193c0f0cd6a9dba3e99e25d5","ref":"refs/heads/master","pushedAt":"2024-05-15T07:41:41.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 39ee3a1e\n\n[Continuations] Add helper commonMaterializable()\n[Continuations] Add one extra SROA\n[Continuations] Add PostProcessing test.\n[Continuations] Check that resume functions report no stack usage\n[Continuations] Cleanup includes.\n[Continuations] Fix corruption seen in RaytracingReflections\n[Continuations] Fix first payload argument dword calculation\n[Continuations] Fix outgoing register count for kernelentry\n[Continuations] Get DispatchSystemData type from GetLocalRootIndex.\n[Continuations] Handle gpurt GetSetting\n[Continuations] Ignore life time markers when doing store-to-load forwarding\n[Continuations] Keep _AmdAcceptHitAttributes calls\n[Continuations] Lower _GetResumePoint correctly\n[Continuations] Move stackptr lowering\n[Continuations] Promote _AmdAwait arguments correctly\n[Continuations] Refactor RT shader stage enum\n[Continuations] Rematerialize single source shuffle\n[Continuations] Remove old unused helper\n[Continuations] Remove use of [[no_unique_address]]\n[Continuations] Set up proper register count for Traversal\n[Continuations] Spill in PayloadCopyHelper\n[Continuations] Support TriangleVertexPositions in continuations\n[Continuations] Update test to reduce diff\n[Continuations] Use int as name for gpurt settings\n[Continuations] Use llvmraytracing in more places\n[RT] Set the entry name as modulename\nAdd 'override' to SPIRVBaseVariable::validate()\nAdd AmdExtD3DShaderIntrinsics_ShaderMarker\nAdd DbgInfo api to ModuleBunch\nAdd detecting more patterns which can be lowered into fmul_legacy and fma_legacy\nAdd dynamicVgprBlockSize option\nAdd GpurtVersionFlags\nAdd Phoenix2 support\nAdd ShaderModuleUsage::usePrimitiveId\nApply WQM to result of WWM subgroup operations\nClean GEP formation for NGG atomic add\nCleanup the old PAL metadata code path\nCrossModuleInliner: add check for consistent target module\nCrossModuleInliner: make struct layout independent of NDEBUG\nDataLayout index sizes for SPIRV address spaces\nDetect read-only buffers and mark them as invariant\nExpand GEP compatibility in tests\nHandle new return type for some IRBuilder methods\nHandle new return type from DIBuilder\nlgc: add strided buffer load\nlgc: correct stage detection for subgroup builder\nlgc: missing code for load.strided.buffer.desc\nlgc: Add LoadBufferAddrOp\nlgc: Allow hoisting common instructions\nlgc: Disassemble archive-of-ELFs\nlgc: move CreateDebugBreak to BuilderCommon\nlgc: Set flat workgroup size for cs_chain\nLowerRayTracingPipeline: Avoid redundant GEPs\nMove replaceAllPointerUses into CompilerUtils\nPack mesh shader outputs in LDS space\nReduce SW XFB LDS usage\nRefine advanced blend func with value-type arguments\nRefine dump info for shader module id\nRefine pipeline dump when shader module identifier is used\nRemove checks for gfxip >= 10\nRemove LDS granularity rounding\nRemove ShaderModuleUsage::useGenericBuiltIn\nRemove the accidentally-added file for *.pipe.swp\nReport new RequireFullPipeline error\nReturn frozen poison value in _AmdGetUninitialized\nReturn the state whether the texture is accessed by texelFetch function\nSet MaxOutgoingVgprCount into SGPR and pass it\nSimplify LDS GEP formation\nSome renaming in mesh shader\nSupport uint64 AmdTraceRaySampleGpuTimer()\nTargetFeatures for amdgpu_cs_chain_preserve funcs\nUpdate llvm-dialects submodule\nUse 32-bit GEPs consistently\nUse CmpInst::Create instead of new to create instruction\nUse convertUsersOfConstantsToInstruction throughout\nVertexFetch: vertex offset is added twice in vertex offset mode\nFix check_llpc test regression\nFix color export offline compilation\nFix compiler warnings\nFix error on unused private field\nFix ShaderDbg issues\nFix the check for reserving a dummy vertex attribute\nFix unusual case imported by new game:STARFIELD\nFix updating threshold issue when merge with pipeline cache which cause PAL assert","shortMessageHtmlLink":"Update llpc from commit 39ee3a1e"}},{"before":"a80cc56409e5c209acdd2a482f846409073b2806","after":"8252cfa83e03aa0a193c0f0cd6a9dba3e99e25d5","ref":"refs/heads/dev","pushedAt":"2024-05-15T07:18:35.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 39ee3a1e\n\n[Continuations] Add helper commonMaterializable()\n[Continuations] Add one extra SROA\n[Continuations] Add PostProcessing test.\n[Continuations] Check that resume functions report no stack usage\n[Continuations] Cleanup includes.\n[Continuations] Fix corruption seen in RaytracingReflections\n[Continuations] Fix first payload argument dword calculation\n[Continuations] Fix outgoing register count for kernelentry\n[Continuations] Get DispatchSystemData type from GetLocalRootIndex.\n[Continuations] Handle gpurt GetSetting\n[Continuations] Ignore life time markers when doing store-to-load forwarding\n[Continuations] Keep _AmdAcceptHitAttributes calls\n[Continuations] Lower _GetResumePoint correctly\n[Continuations] Move stackptr lowering\n[Continuations] Promote _AmdAwait arguments correctly\n[Continuations] Refactor RT shader stage enum\n[Continuations] Rematerialize single source shuffle\n[Continuations] Remove old unused helper\n[Continuations] Remove use of [[no_unique_address]]\n[Continuations] Set up proper register count for Traversal\n[Continuations] Spill in PayloadCopyHelper\n[Continuations] Support TriangleVertexPositions in continuations\n[Continuations] Update test to reduce diff\n[Continuations] Use int as name for gpurt settings\n[Continuations] Use llvmraytracing in more places\n[RT] Set the entry name as modulename\nAdd 'override' to SPIRVBaseVariable::validate()\nAdd AmdExtD3DShaderIntrinsics_ShaderMarker\nAdd DbgInfo api to ModuleBunch\nAdd detecting more patterns which can be lowered into fmul_legacy and fma_legacy\nAdd dynamicVgprBlockSize option\nAdd GpurtVersionFlags\nAdd Phoenix2 support\nAdd ShaderModuleUsage::usePrimitiveId\nApply WQM to result of WWM subgroup operations\nClean GEP formation for NGG atomic add\nCleanup the old PAL metadata code path\nCrossModuleInliner: add check for consistent target module\nCrossModuleInliner: make struct layout independent of NDEBUG\nDataLayout index sizes for SPIRV address spaces\nDetect read-only buffers and mark them as invariant\nExpand GEP compatibility in tests\nHandle new return type for some IRBuilder methods\nHandle new return type from DIBuilder\nlgc: add strided buffer load\nlgc: correct stage detection for subgroup builder\nlgc: missing code for load.strided.buffer.desc\nlgc: Add LoadBufferAddrOp\nlgc: Allow hoisting common instructions\nlgc: Disassemble archive-of-ELFs\nlgc: move CreateDebugBreak to BuilderCommon\nlgc: Set flat workgroup size for cs_chain\nLowerRayTracingPipeline: Avoid redundant GEPs\nMove replaceAllPointerUses into CompilerUtils\nPack mesh shader outputs in LDS space\nReduce SW XFB LDS usage\nRefine advanced blend func with value-type arguments\nRefine dump info for shader module id\nRefine pipeline dump when shader module identifier is used\nRemove checks for gfxip >= 10\nRemove LDS granularity rounding\nRemove ShaderModuleUsage::useGenericBuiltIn\nRemove the accidentally-added file for *.pipe.swp\nReport new RequireFullPipeline error\nReturn frozen poison value in _AmdGetUninitialized\nReturn the state whether the texture is accessed by texelFetch function\nSet MaxOutgoingVgprCount into SGPR and pass it\nSimplify LDS GEP formation\nSome renaming in mesh shader\nSupport uint64 AmdTraceRaySampleGpuTimer()\nTargetFeatures for amdgpu_cs_chain_preserve funcs\nUpdate llvm-dialects submodule\nUse 32-bit GEPs consistently\nUse CmpInst::Create instead of new to create instruction\nUse convertUsersOfConstantsToInstruction throughout\nVertexFetch: vertex offset is added twice in vertex offset mode\nFix check_llpc test regression\nFix color export offline compilation\nFix compiler warnings\nFix error on unused private field\nFix ShaderDbg issues\nFix the check for reserving a dummy vertex attribute\nFix unusual case imported by new game:STARFIELD\nFix updating threshold issue when merge with pipeline cache which cause PAL assert","shortMessageHtmlLink":"Update llpc from commit 39ee3a1e"}},{"before":"fba35920322463e9781b82c9f0016f06d904dfee","after":"3f33e94c6faf594537023ccbdafc0ca45dc0315d","ref":"refs/heads/dev_promote","pushedAt":"2024-05-15T03:43:14.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix static check error","shortMessageHtmlLink":"Fix static check error"}},{"before":"1d70e90038eb367d6a8ad5f11a2404efd9e200cf","after":"fba35920322463e9781b82c9f0016f06d904dfee","ref":"refs/heads/dev_promote","pushedAt":"2024-05-14T10:21:43.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Install python3-ruamel.yaml","shortMessageHtmlLink":"Install python3-ruamel.yaml"}},{"before":"f875b934a6fc6f0684e63124610e0a01aeed8f45","after":"1d70e90038eb367d6a8ad5f11a2404efd9e200cf","ref":"refs/heads/dev_promote","pushedAt":"2024-05-14T10:14:18.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix python3-ruamel.yaml install","shortMessageHtmlLink":"Fix python3-ruamel.yaml install"}},{"before":"266215c3faa0d298fc77c7ed64eb1eff82780946","after":"f875b934a6fc6f0684e63124610e0a01aeed8f45","ref":"refs/heads/dev_promote","pushedAt":"2024-05-14T10:09:25.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Install python3-ruamel.yaml","shortMessageHtmlLink":"Install python3-ruamel.yaml"}},{"before":"14535c028dc70a093d95a95efedb749812c1e77d","after":"a80cc56409e5c209acdd2a482f846409073b2806","ref":"refs/heads/dev","pushedAt":"2024-05-14T09:54:32.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"[CI] Install python3-ruamel.yaml","shortMessageHtmlLink":"[CI] Install python3-ruamel.yaml"}},{"before":"14535c028dc70a093d95a95efedb749812c1e77d","after":"266215c3faa0d298fc77c7ed64eb1eff82780946","ref":"refs/heads/dev_promote","pushedAt":"2024-05-14T09:38:07.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit d6557077\n\n[Continuations] Add helper commonMaterializable()\n[Continuations] Add one extra SROA\n[Continuations] Add PostProcessing test.\n[Continuations] Check that resume functions report no stack usage\n[Continuations] Cleanup includes.\n[Continuations] Fix corruption seen in RaytracingReflections\n[Continuations] Fix first payload argument dword calculation\n[Continuations] Fix outgoing register count for kernelentry\n[Continuations] Get DispatchSystemData type from GetLocalRootIndex.\n[Continuations] Handle gpurt GetSetting\n[Continuations] Ignore life time markers when doing store-to-load forwarding\n[Continuations] Keep _AmdAcceptHitAttributes calls\n[Continuations] Lower _GetResumePoint correctly\n[Continuations] Move stackptr lowering\n[Continuations] Promote _AmdAwait arguments correctly\n[Continuations] Refactor RT shader stage enum\n[Continuations] Rematerialize single source shuffle\n[Continuations] Remove old unused helper\n[Continuations] Remove use of [[no_unique_address]]\n[Continuations] Set up proper register count for Traversal\n[Continuations] Spill in PayloadCopyHelper\n[Continuations] Support TriangleVertexPositions in continuations\n[Continuations] Update test to reduce diff\n[Continuations] Use int as name for gpurt settings\n[Continuations] Use llvmraytracing in more places\n[RT] Set the entry name as modulename\nAdd 'override' to SPIRVBaseVariable::validate()\nAdd AmdExtD3DShaderIntrinsics_ShaderMarker\nAdd DbgInfo api to ModuleBunch\nAdd detecting more patterns which can be lowered into fmul_legacy and fma_legacy\nAdd dynamicVgprBlockSize option\nAdd GpurtVersionFlags\nAdd Phoenix2 support\nAdd ShaderModuleUsage::usePrimitiveId\nApply WQM to result of WWM subgroup operations\nClean GEP formation for NGG atomic add\nCleanup the old PAL metadata code path\nCrossModuleInliner: add check for consistent target module\nCrossModuleInliner: make struct layout independent of NDEBUG\nDataLayout index sizes for SPIRV address spaces\nDetect read-only buffers and mark them as invariant\nExpand GEP compatibility in tests\nHandle new return type for some IRBuilder methods\nHandle new return type from DIBuilder\nlgc: add strided buffer load\nlgc: correct stage detection for subgroup builder\nlgc: missing code for load.strided.buffer.desc\nlgc: Add LoadBufferAddrOp\nlgc: Allow hoisting common instructions\nlgc: Disassemble archive-of-ELFs\nlgc: move CreateDebugBreak to BuilderCommon\nlgc: Set flat workgroup size for cs_chain\nlgc: Unbreak dxcp build\nLowerRayTracingPipeline: Avoid redundant GEPs\nMove replaceAllPointerUses into CompilerUtils\nPack mesh shader outputs in LDS space\nReduce SW XFB LDS usage\nRefine advanced blend func with value-type arguments\nRefine dump info for shader module id\nRefine pipeline dump when shader module identifier is used\nRemove checks for gfxip >= 10\nRemove LDS granularity rounding\nRemove ShaderModuleUsage::useGenericBuiltIn\nRemove the accidentally-added file for *.pipe.swp\nReport new RequireFullPipeline error\nReturn frozen poison value in _AmdGetUninitialized\nReturn the state whether the texture is accessed by texelFetch function\nSet MaxOutgoingVgprCount into SGPR and pass it\nSimplify LDS GEP formation\nSome renaming in mesh shader\nSupport uint64 AmdTraceRaySampleGpuTimer()\nTargetFeatures for amdgpu_cs_chain_preserve funcs\nUpdate llvm-dialects submodule\nUse 32-bit GEPs consistently\nUse CmpInst::Create instead of new to create instruction\nUse convertUsersOfConstantsToInstruction throughout\nVertexFetch: vertex offset is added twice in vertex offset mode\nFix check_llpc test regression\nFix color export offline compilation\nFix compiler warnings\nFix error on unused private field\nFix ShaderDbg issues\nFix the check for reserving a dummy vertex attribute\nFix unusual case imported by new game:STARFIELD\nFix updating threshold issue when merge with pipeline cache which cause PAL assert","shortMessageHtmlLink":"Update llpc from commit d6557077"}},{"before":"d4c8845341ffa6cf0e20b7c805a96e13055a7164","after":"14535c028dc70a093d95a95efedb749812c1e77d","ref":"refs/heads/master","pushedAt":"2024-03-26T04:22:59.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 626bce4e\n\nlgc: Add ShaderDbg support for ElfLinker\nlgc: Cooperative rowacc code clean up\nlgc: finish builder-ification of PatchInOutImportExport\nlgc: Fix cooperative matrix enum build error\nlgc: Move cooperative matrix enum into lgcDialect\nlgc: SamplePosOffset should be v2f32 instead of i32\nlgc: Shrink down include list.\nlgc: Support waveMatrix rowacc expand\nlgc: Support waveMatrix rowacc scalar\nlgc: Support waveMatrix rowacc splat (fill)\nlgc: Support waveMatrix rowacc sum accumulate\nlgc: Treat patch + GS w/o tess to be point mode\nlgc: Undo moving code into LowerSubgroupOps\nlgc: Use llvm-dialects with new name mangling\nllpc: fix a corner case with LDS in task/mesh pipelines\n[Continuations] Add _AmdContStackLoadLastUse builtin support.\n[Continuations] Add _cont_-prefixed contants.\n[Continuations] Do not change function name during linking\n[Continuations] Do not remove types metadata\n[Continuations] Fix continuations-units test warning.\n[Continuations] Fix undef / poison test changes.\n[Continuations] Fix use-after-free\n[Continuations] Handle InstanceInclusionMaskOp\n[Continuations] Handle mangled _AmdComplete calls.\n[Continuations] Implement lgc.cps lowering in DXILContPostProcess.\n[Continuations] Lower GetResumePoint\n[Continuations] Pass GPURT library to LowerRaytracingPipelinePass\n[Continuations] Process lgc.cps function with no await call correctly\n[Continuations] Refactor DXILContIntrinsicPrepare\n[Continuations] Refactor pipeline-scope payload size handling\n[Continuations] Remove continuation.complete intrinsic.\n[Continuations] Remove more deprecated 'startswith' calls.\n[Continuations] Remove types metadata sometimes\n[Continuations] Replace payload spill allocation with alloca.\n[Continuations] Terminate shader after inlining _cont_SetupRayGen.\n[RT] Bad code on calculating hash for RT pipeline which contains library\n[RT] Fix tracedata for function return\n[RT] Remove all unnecessary functions for Traversal\n[RT] Replace deprecated startswith call.\n[RT] Swap InstanceIndexOp and InstanceIdOp\n[RT] Use more GpuRt Funcs instead of hardcoded ir.\n[RT] use the getinstance index/id from gpurt directly.\nAbort for -emit-llvm and color export shader\nAdd new field 'numComponents' to 'InOutInfo'\nAdd td file in cmake source files\nAdd the alignment for cooperativeMatrixLoad and cooperativeMatrixStore\nChanges for non-intrisic debug info mode\nCleanup-gfx9: Remove check for major >= 9\nCorrect the condition for executing color export operation\nMove shared/continuations/ to llvmraytracing/\nPort the optimization for the pattern: ((b==0.0 ? 0.0 : a) * (a==0.0 ? 0.0 : b))\nPreserve necessary _Amd function calls\nrayquery: Add first version of lgc.rtq dialect\nRefactor handleColorExportInstructions\nRefine VKGC pipeline dump and vk sections in vfx\nShaderStages test check lines more flexible\nSupport PipelineShaderOptions::clientHash in pipeline hash calculation\nUpdate llpc tests to allow for upstream llvm changes\nUpdate to handle changes in PassModel parameters\nUse VKI_ macros in llpc_version\nWrap PipelineContext and RaytracingContext in SPIRVReader.\nBump Gpurt Version to 45\nUpdate llvm-dialects.\nFix Attributes for Cooperative Matrices\nFix fp and packed formats' norm emulation issue.\nFix RT demo perf-drop due to handling global access for null desc\nFix some CTS failures for dynamicDualSource on NAVI10","shortMessageHtmlLink":"Update llpc from commit 626bce4e"}},{"before":"e91a935d9e3ae526f4cd8044659609ba8daa858b","after":"14535c028dc70a093d95a95efedb749812c1e77d","ref":"refs/heads/dev","pushedAt":"2024-03-26T03:39:32.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 626bce4e\n\nlgc: Add ShaderDbg support for ElfLinker\nlgc: Cooperative rowacc code clean up\nlgc: finish builder-ification of PatchInOutImportExport\nlgc: Fix cooperative matrix enum build error\nlgc: Move cooperative matrix enum into lgcDialect\nlgc: SamplePosOffset should be v2f32 instead of i32\nlgc: Shrink down include list.\nlgc: Support waveMatrix rowacc expand\nlgc: Support waveMatrix rowacc scalar\nlgc: Support waveMatrix rowacc splat (fill)\nlgc: Support waveMatrix rowacc sum accumulate\nlgc: Treat patch + GS w/o tess to be point mode\nlgc: Undo moving code into LowerSubgroupOps\nlgc: Use llvm-dialects with new name mangling\nllpc: fix a corner case with LDS in task/mesh pipelines\n[Continuations] Add _AmdContStackLoadLastUse builtin support.\n[Continuations] Add _cont_-prefixed contants.\n[Continuations] Do not change function name during linking\n[Continuations] Do not remove types metadata\n[Continuations] Fix continuations-units test warning.\n[Continuations] Fix undef / poison test changes.\n[Continuations] Fix use-after-free\n[Continuations] Handle InstanceInclusionMaskOp\n[Continuations] Handle mangled _AmdComplete calls.\n[Continuations] Implement lgc.cps lowering in DXILContPostProcess.\n[Continuations] Lower GetResumePoint\n[Continuations] Pass GPURT library to LowerRaytracingPipelinePass\n[Continuations] Process lgc.cps function with no await call correctly\n[Continuations] Refactor DXILContIntrinsicPrepare\n[Continuations] Refactor pipeline-scope payload size handling\n[Continuations] Remove continuation.complete intrinsic.\n[Continuations] Remove more deprecated 'startswith' calls.\n[Continuations] Remove types metadata sometimes\n[Continuations] Replace payload spill allocation with alloca.\n[Continuations] Terminate shader after inlining _cont_SetupRayGen.\n[RT] Bad code on calculating hash for RT pipeline which contains library\n[RT] Fix tracedata for function return\n[RT] Remove all unnecessary functions for Traversal\n[RT] Replace deprecated startswith call.\n[RT] Swap InstanceIndexOp and InstanceIdOp\n[RT] Use more GpuRt Funcs instead of hardcoded ir.\n[RT] use the getinstance index/id from gpurt directly.\nAbort for -emit-llvm and color export shader\nAdd new field 'numComponents' to 'InOutInfo'\nAdd td file in cmake source files\nAdd the alignment for cooperativeMatrixLoad and cooperativeMatrixStore\nChanges for non-intrisic debug info mode\nCleanup-gfx9: Remove check for major >= 9\nCorrect the condition for executing color export operation\nMove shared/continuations/ to llvmraytracing/\nPort the optimization for the pattern: ((b==0.0 ? 0.0 : a) * (a==0.0 ? 0.0 : b))\nPreserve necessary _Amd function calls\nrayquery: Add first version of lgc.rtq dialect\nRefactor handleColorExportInstructions\nRefine VKGC pipeline dump and vk sections in vfx\nShaderStages test check lines more flexible\nSupport PipelineShaderOptions::clientHash in pipeline hash calculation\nUpdate llpc tests to allow for upstream llvm changes\nUpdate to handle changes in PassModel parameters\nUse VKI_ macros in llpc_version\nWrap PipelineContext and RaytracingContext in SPIRVReader.\nBump Gpurt Version to 45\nUpdate llvm-dialects.\nFix Attributes for Cooperative Matrices\nFix fp and packed formats' norm emulation issue.\nFix RT demo perf-drop due to handling global access for null desc\nFix some CTS failures for dynamicDualSource on NAVI10","shortMessageHtmlLink":"Update llpc from commit 626bce4e"}},{"before":"b1f609e6cb38ef1db6bbd3a87caa4c09f080cb3a","after":"14535c028dc70a093d95a95efedb749812c1e77d","ref":"refs/heads/dev_promote","pushedAt":"2024-03-26T03:34:37.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit 626bce4e\n\nlgc: Add ShaderDbg support for ElfLinker\nlgc: Cooperative rowacc code clean up\nlgc: finish builder-ification of PatchInOutImportExport\nlgc: Fix cooperative matrix enum build error\nlgc: Move cooperative matrix enum into lgcDialect\nlgc: SamplePosOffset should be v2f32 instead of i32\nlgc: Shrink down include list.\nlgc: Support waveMatrix rowacc expand\nlgc: Support waveMatrix rowacc scalar\nlgc: Support waveMatrix rowacc splat (fill)\nlgc: Support waveMatrix rowacc sum accumulate\nlgc: Treat patch + GS w/o tess to be point mode\nlgc: Undo moving code into LowerSubgroupOps\nlgc: Use llvm-dialects with new name mangling\nllpc: fix a corner case with LDS in task/mesh pipelines\n[Continuations] Add _AmdContStackLoadLastUse builtin support.\n[Continuations] Add _cont_-prefixed contants.\n[Continuations] Do not change function name during linking\n[Continuations] Do not remove types metadata\n[Continuations] Fix continuations-units test warning.\n[Continuations] Fix undef / poison test changes.\n[Continuations] Fix use-after-free\n[Continuations] Handle InstanceInclusionMaskOp\n[Continuations] Handle mangled _AmdComplete calls.\n[Continuations] Implement lgc.cps lowering in DXILContPostProcess.\n[Continuations] Lower GetResumePoint\n[Continuations] Pass GPURT library to LowerRaytracingPipelinePass\n[Continuations] Process lgc.cps function with no await call correctly\n[Continuations] Refactor DXILContIntrinsicPrepare\n[Continuations] Refactor pipeline-scope payload size handling\n[Continuations] Remove continuation.complete intrinsic.\n[Continuations] Remove more deprecated 'startswith' calls.\n[Continuations] Remove types metadata sometimes\n[Continuations] Replace payload spill allocation with alloca.\n[Continuations] Terminate shader after inlining _cont_SetupRayGen.\n[RT] Bad code on calculating hash for RT pipeline which contains library\n[RT] Fix tracedata for function return\n[RT] Remove all unnecessary functions for Traversal\n[RT] Replace deprecated startswith call.\n[RT] Swap InstanceIndexOp and InstanceIdOp\n[RT] Use more GpuRt Funcs instead of hardcoded ir.\n[RT] use the getinstance index/id from gpurt directly.\nAbort for -emit-llvm and color export shader\nAdd new field 'numComponents' to 'InOutInfo'\nAdd td file in cmake source files\nAdd the alignment for cooperativeMatrixLoad and cooperativeMatrixStore\nChanges for non-intrisic debug info mode\nCleanup-gfx9: Remove check for major >= 9\nCorrect the condition for executing color export operation\nMove shared/continuations/ to llvmraytracing/\nPort the optimization for the pattern: ((b==0.0 ? 0.0 : a) * (a==0.0 ? 0.0 : b))\nPreserve necessary _Amd function calls\nrayquery: Add first version of lgc.rtq dialect\nRefactor handleColorExportInstructions\nRefine VKGC pipeline dump and vk sections in vfx\nShaderStages test check lines more flexible\nSupport PipelineShaderOptions::clientHash in pipeline hash calculation\nUpdate llpc tests to allow for upstream llvm changes\nUpdate to handle changes in PassModel parameters\nUse VKI_ macros in llpc_version\nWrap PipelineContext and RaytracingContext in SPIRVReader.\nBump Gpurt Version to 45\nUpdate llvm-dialects.\nFix Attributes for Cooperative Matrices\nFix fp and packed formats' norm emulation issue.\nFix RT demo perf-drop due to handling global access for null desc\nFix some CTS failures for dynamicDualSource on NAVI10","shortMessageHtmlLink":"Update llpc from commit 626bce4e"}},{"before":"fe14fbb74847f805c4ee583dad4aad5753313069","after":"b1f609e6cb38ef1db6bbd3a87caa4c09f080cb3a","ref":"refs/heads/dev_promote","pushedAt":"2024-03-22T10:08:30.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix cpp static check (2) update.","shortMessageHtmlLink":"Fix cpp static check (2) update."}},{"before":"0f4d3fe538a58bb65def97c88bd05016659d2745","after":"fe14fbb74847f805c4ee583dad4aad5753313069","ref":"refs/heads/dev_promote","pushedAt":"2024-03-22T07:02:38.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix cpp static check.","shortMessageHtmlLink":"Fix cpp static check."}},{"before":"65196b5be9e1f3ee1c322d0987ed94b06c4d3c20","after":"e91a935d9e3ae526f4cd8044659609ba8daa858b","ref":"refs/heads/dev","pushedAt":"2024-03-22T03:17:43.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit f43ed3d0\n\nlgc: Add ShaderDbg support for ElfLinker\nlgc: Cooperative rowacc code clean up\nlgc: finish builder-ification of PatchInOutImportExport\nlgc: Fix cooperative matrix enum build error\nlgc: Move cooperative matrix enum into lgcDialect\nlgc: SamplePosOffset should be v2f32 instead of i32\nlgc: Shrink down include list.\nlgc: Support waveMatrix rowacc expand\nlgc: Support waveMatrix rowacc scalar\nlgc: Support waveMatrix rowacc splat (fill)\nlgc: Support waveMatrix rowacc sum accumulate\nlgc: Treat patch + GS w/o tess to be point mode\nlgc: Undo moving code into LowerSubgroupOps\nlgc: Use llvm-dialects with new name mangling\nllpc: fix a corner case with LDS in task/mesh pipelines\n[Continuations] Add _AmdContStackLoadLastUse builtin support.\n[Continuations] Add _cont_-prefixed contants.\n[Continuations] Do not change function name during linking\n[Continuations] Do not remove types metadata\n[Continuations] Fix continuations-units test warning.\n[Continuations] Fix undef / poison test changes.\n[Continuations] Fix use-after-free\n[Continuations] Handle InstanceInclusionMaskOp\n[Continuations] Handle mangled _AmdComplete calls.\n[Continuations] Implement lgc.cps lowering in DXILContPostProcess.\n[Continuations] Lower GetResumePoint\n[Continuations] Pass GPURT library to LowerRaytracingPipelinePass\n[Continuations] Process lgc.cps function with no await call correctly\n[Continuations] Refactor DXILContIntrinsicPrepare\n[Continuations] Refactor pipeline-scope payload size handling\n[Continuations] Remove continuation.complete intrinsic.\n[Continuations] Remove more deprecated 'startswith' calls.\n[Continuations] Remove types metadata sometimes\n[Continuations] Replace payload spill allocation with alloca.\n[Continuations] Terminate shader after inlining _cont_SetupRayGen.\n[RT] Bad code on calculating hash for RT pipeline which contains library\n[RT] Fix tracedata for function return\n[RT] Remove all unnecessary functions for Traversal\n[RT] Replace deprecated startswith call.\n[RT] Swap InstanceIndexOp and InstanceIdOp\n[RT] Use more GpuRt Funcs instead of hardcoded ir.\n[RT] use the getinstance index/id from gpurt directly.\nAbort for -emit-llvm and color export shader\nAdd new field 'numComponents' to 'InOutInfo'\nAdd td file in cmake source files\nAdd the alignment for cooperativeMatrixLoad and cooperativeMatrixStore\nChanges for non-intrisic debug info mode\nCleanup-gfx9: Remove check for major >= 9\nCorrect the condition for executing color export operation\nMove shared/continuations/ to llvmraytracing/\nPort the optimization for the pattern: ((b==0.0 ? 0.0 : a) * (a==0.0 ? 0.0 : b))\nPreserve necessary _Amd function calls\nrayquery: Add first version of lgc.rtq dialect\nRefactor handleColorExportInstructions\nRefine VKGC pipeline dump and vk sections in vfx\nShaderStages test check lines more flexible\nSupport PipelineShaderOptions::clientHash in pipeline hash calculation\nUpdate llpc tests to allow for upstream llvm changes\nUpdate to handle changes in PassModel parameters\nUse VKI_ macros in llpc_version\nWrap PipelineContext and RaytracingContext in SPIRVReader.\nBump Gpurt Version to 45\nUpdate llvm-dialects.\nFix Attributes for Cooperative Matrices\nFix fp and packed formats' norm emulation issue.\nFix RT demo perf-drop due to handling global access for null desc\nFix some CTS failures for dynamicDualSource on NAVI10","shortMessageHtmlLink":"Update llpc from commit f43ed3d0"}},{"before":"e91a935d9e3ae526f4cd8044659609ba8daa858b","after":"0f4d3fe538a58bb65def97c88bd05016659d2745","ref":"refs/heads/dev_promote","pushedAt":"2024-03-22T03:10:20.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix llpc update error","shortMessageHtmlLink":"Fix llpc update error"}},{"before":"23fdc73ec66db47f76e54d26c146e9d904f5aab4","after":"e91a935d9e3ae526f4cd8044659609ba8daa858b","ref":"refs/heads/dev_promote","pushedAt":"2024-03-22T02:33:43.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit f43ed3d0\n\nlgc: Add ShaderDbg support for ElfLinker\nlgc: Cooperative rowacc code clean up\nlgc: finish builder-ification of PatchInOutImportExport\nlgc: Fix cooperative matrix enum build error\nlgc: Move cooperative matrix enum into lgcDialect\nlgc: SamplePosOffset should be v2f32 instead of i32\nlgc: Shrink down include list.\nlgc: Support waveMatrix rowacc expand\nlgc: Support waveMatrix rowacc scalar\nlgc: Support waveMatrix rowacc splat (fill)\nlgc: Support waveMatrix rowacc sum accumulate\nlgc: Treat patch + GS w/o tess to be point mode\nlgc: Undo moving code into LowerSubgroupOps\nlgc: Use llvm-dialects with new name mangling\nllpc: fix a corner case with LDS in task/mesh pipelines\n[Continuations] Add _AmdContStackLoadLastUse builtin support.\n[Continuations] Add _cont_-prefixed contants.\n[Continuations] Do not change function name during linking\n[Continuations] Do not remove types metadata\n[Continuations] Fix continuations-units test warning.\n[Continuations] Fix undef / poison test changes.\n[Continuations] Fix use-after-free\n[Continuations] Handle InstanceInclusionMaskOp\n[Continuations] Handle mangled _AmdComplete calls.\n[Continuations] Implement lgc.cps lowering in DXILContPostProcess.\n[Continuations] Lower GetResumePoint\n[Continuations] Pass GPURT library to LowerRaytracingPipelinePass\n[Continuations] Process lgc.cps function with no await call correctly\n[Continuations] Refactor DXILContIntrinsicPrepare\n[Continuations] Refactor pipeline-scope payload size handling\n[Continuations] Remove continuation.complete intrinsic.\n[Continuations] Remove more deprecated 'startswith' calls.\n[Continuations] Remove types metadata sometimes\n[Continuations] Replace payload spill allocation with alloca.\n[Continuations] Terminate shader after inlining _cont_SetupRayGen.\n[RT] Bad code on calculating hash for RT pipeline which contains library\n[RT] Fix tracedata for function return\n[RT] Remove all unnecessary functions for Traversal\n[RT] Replace deprecated startswith call.\n[RT] Swap InstanceIndexOp and InstanceIdOp\n[RT] Use more GpuRt Funcs instead of hardcoded ir.\n[RT] use the getinstance index/id from gpurt directly.\nAbort for -emit-llvm and color export shader\nAdd new field 'numComponents' to 'InOutInfo'\nAdd td file in cmake source files\nAdd the alignment for cooperativeMatrixLoad and cooperativeMatrixStore\nChanges for non-intrisic debug info mode\nCleanup-gfx9: Remove check for major >= 9\nCorrect the condition for executing color export operation\nMove shared/continuations/ to llvmraytracing/\nPort the optimization for the pattern: ((b==0.0 ? 0.0 : a) * (a==0.0 ? 0.0 : b))\nPreserve necessary _Amd function calls\nrayquery: Add first version of lgc.rtq dialect\nRefactor handleColorExportInstructions\nRefine VKGC pipeline dump and vk sections in vfx\nShaderStages test check lines more flexible\nSupport PipelineShaderOptions::clientHash in pipeline hash calculation\nUpdate llpc tests to allow for upstream llvm changes\nUpdate to handle changes in PassModel parameters\nUse VKI_ macros in llpc_version\nWrap PipelineContext and RaytracingContext in SPIRVReader.\nBump Gpurt Version to 45\nUpdate llvm-dialects.\nFix Attributes for Cooperative Matrices\nFix fp and packed formats' norm emulation issue.\nFix RT demo perf-drop due to handling global access for null desc\nFix some CTS failures for dynamicDualSource on NAVI10","shortMessageHtmlLink":"Update llpc from commit f43ed3d0"}},{"before":"16f50cd7667ff2bc91e30f66ec178a8e14c97688","after":"23fdc73ec66db47f76e54d26c146e9d904f5aab4","ref":"refs/heads/dev_promote","pushedAt":"2024-03-21T10:40:19.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Fix cpp static check.","shortMessageHtmlLink":"Fix cpp static check."}},{"before":"5579b5e4cafe9804def5dc8389b731846f2a28da","after":"16f50cd7667ff2bc91e30f66ec178a8e14c97688","ref":"refs/heads/dev_promote","pushedAt":"2024-03-21T10:22:58.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit a04cfb48\n\nlgc: Add ShaderDbg support for ElfLinker\nlgc: Cooperative rowacc code clean up\nlgc: finish builder-ification of PatchInOutImportExport\nlgc: Fix cooperative matrix enum build error\nlgc: Move cooperative matrix enum into lgcDialect\nlgc: SamplePosOffset should be v2f32 instead of i32\nlgc: Shrink down include list.\nlgc: Support waveMatrix rowacc expand\nlgc: Support waveMatrix rowacc scalar\nlgc: Support waveMatrix rowacc splat (fill)\nlgc: Support waveMatrix rowacc sum accumulate\nlgc: Treat patch + GS w/o tess to be point mode\nlgc: Undo moving code into LowerSubgroupOps\nlgc: Use llvm-dialects with new name mangling\nllpc: fix a corner case with LDS in task/mesh pipelines\n[Continuations] Add _AmdContStackLoadLastUse builtin support.\n[Continuations] Add _cont_-prefixed contants.\n[Continuations] Do not change function name during linking\n[Continuations] Do not remove types metadata\n[Continuations] Fix continuations-units test warning.\n[Continuations] Fix undef / poison test changes.\n[Continuations] Fix use-after-free\n[Continuations] Handle InstanceInclusionMaskOp\n[Continuations] Handle mangled _AmdComplete calls.\n[Continuations] Implement lgc.cps lowering in DXILContPostProcess.\n[Continuations] Lower GetResumePoint\n[Continuations] Pass GPURT library to LowerRaytracingPipelinePass\n[Continuations] Process lgc.cps function with no await call correctly\n[Continuations] Refactor DXILContIntrinsicPrepare\n[Continuations] Refactor pipeline-scope payload size handling\n[Continuations] Remove continuation.complete intrinsic.\n[Continuations] Remove more deprecated 'startswith' calls.\n[Continuations] Remove types metadata sometimes\n[Continuations] Replace payload spill allocation with alloca.\n[Continuations] Terminate shader after inlining _cont_SetupRayGen.\n[RT] Bad code on calculating hash for RT pipeline which contains library\n[RT] Fix tracedata for function return\n[RT] Remove all unnecessary functions for Traversal\n[RT] Replace deprecated startswith call.\n[RT] Swap InstanceIndexOp and InstanceIdOp\n[RT] Use more GpuRt Funcs instead of hardcoded ir.\n[RT] use the getinstance index/id from gpurt directly.\nAbort for -emit-llvm and color export shader\nAdd new field 'numComponents' to 'InOutInfo'\nAdd td file in cmake source files\nAdd the alignment for cooperativeMatrixLoad and cooperativeMatrixStore\nChanges for non-intrisic debug info mode\nCleanup-gfx9: Remove check for major >= 9\nCorrect the condition for executing color export operation\nMove shared/continuations/ to llvmraytracing/\nPort the optimization for the pattern: ((b==0.0 ? 0.0 : a) * (a==0.0 ? 0.0 : b))\nPreserve necessary _Amd function calls\nrayquery: Add first version of lgc.rtq dialect\nRefactor handleColorExportInstructions\nRefine VKGC pipeline dump and vk sections in vfx\nShaderStages test check lines more flexible\nSupport PipelineShaderOptions::clientHash in pipeline hash calculation\nUpdate llpc tests to allow for upstream llvm changes\nUpdate to handle changes in PassModel parameters\nUse VKI_ macros in llpc_version\nWrap PipelineContext and RaytracingContext in SPIRVReader.\nBump Gpurt Version to 45\nUpdate llvm-dialects.\nFix Attributes for Cooperative Matrices\nFix fp and packed formats' norm emulation issue.\nFix RT demo perf-drop due to handling global access for null desc\nFix some CTS failures for dynamicDualSource on NAVI10","shortMessageHtmlLink":"Update llpc from commit a04cfb48"}},{"before":"eadd37ec3e67cd4ee8b1df2a628c8dcfeca3dfc6","after":"65196b5be9e1f3ee1c322d0987ed94b06c4d3c20","ref":"refs/heads/dev","pushedAt":"2024-02-22T11:02:29.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Promotion from AMD internal branch for 2024.Q1.2","shortMessageHtmlLink":"Promotion from AMD internal branch for 2024.Q1.2"}},{"before":"5579b5e4cafe9804def5dc8389b731846f2a28da","after":"d4c8845341ffa6cf0e20b7c805a96e13055a7164","ref":"refs/heads/master","pushedAt":"2024-02-22T10:59:34.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Update llpc from commit a1369c61\n\n[Continuations] Handle ShaderRecordBufferOp\n[Continuations] Integrate the KernelEntry Function\n[LGC] Add cooperative row acc load and store\n[LGC] Add new pass for lowering of subgroup ops\n[LGC] Add subgroupElect and subgroupAny to dialect\n[RT] Use strided buffer pointers\nCache color export shader\nClean up translation of binary opcodes. NFC.\nConsistently use snake case for dialect op in/out\nCrossModuleInliner: handle the case of incomplete IR\nExpose maximal reconvergence\nPatchBufferOp: Fix the issue of returning null for atomic instructions\nRefine the arguments of SpirvLower::addPasses\nRemove a check of location offset in CreateReadPerVertexInput\nRemove dead code for major < 10\nRemove -enable-scratch-bounds-checks in tests\nRemove fetch shader\nRemove the gating macro for KHR_SHADER_FLOAT_CONTROLS2\nRemove the runImpl pattern almost everywhere\nRemove unnecessary assert in PipelineState::getVerticesPerPrimitive\nRevise two parameter names of markGenericInputOutputUsage\nSupport AmdExtDispatchThreadIdFlat\nSupport pipeline hash specified by driver\nSupport single dword address in uber-fetch shader constant buffer\nUpdate affected llpc/lgc tests as getelementptr now uses i8 instead of a type in certain cases\nUpdate gfx900-based lgc tests to gfx1010\nUpdate SPIR-V header\nUse export format to calculate cache hash for CES\nVK_KHR_shader_expect_assume implementation\nVK_KHR_shader_quad_control implementation\nVK_KHR_shader_subgroup_rotate implementation","shortMessageHtmlLink":"Update llpc from commit a1369c61"}},{"before":"5de2a348b6ac131ecef2e3a8863f0e33609c182a","after":"eadd37ec3e67cd4ee8b1df2a628c8dcfeca3dfc6","ref":"refs/heads/dev","pushedAt":"2024-02-22T10:44:43.000Z","pushType":"push","commitsCount":2,"pusher":{"login":"qiaojbao","name":"Qiaojin.Bao","path":"/qiaojbao","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/52525627?s=80&v=4"},"commit":{"message":"Promotion from AMD internal branch for 2024.Q1.2","shortMessageHtmlLink":"Promotion from AMD internal branch for 2024.Q1.2"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEb3ft9wA","startCursor":null,"endCursor":null}},"title":"Activity · GPUOpen-Drivers/llpc"}