Promotion from AMD internal branch for 2024.Q4.1 #2908
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Set SLC=0 for ATM LowerGpuRt: fix a type confusion Honor NonUniform decorations on OpAccessChain operands Force WGP when NGG is in passthrough mode Use idxen modifier for (RW)StructuredBuffers Compatible with LLVM upstream change that moved Utility funcs to Utils Include Line Correlation Info for Tools via LLPC [Continuations] Cleanup Complete op lowering [CompilerUtils] Improve handling of freeze in ValueOriginTracker [compilerutils] Add GetGlobalInModule to CrossModuleInliner amdllpc: some usability improvements [Continuations] Add SpecializeDriverShadersPass Split up gl_in array type. [Continuations] Remove stack lowering define Use inbounds gep to index row major matrix llvmraytracing: Remove support for _AmdContinuationsGetFlags Properly encapsulate state that affects GPURT specialization llvmraytracing: only support a waitmask of -1 llvmraytracing: remove remaining traces of EnqueueCall compilerutils: fix a warning [LowerGlobals] Mark globals with `buffer.index` users as readonly Implement structural GEP dialect for in/out llvmraytracing: add LLPC/LLVM scratch/global address spaces llpc/ProcessGpuRtLibrary: use earlyGpurtTransform [Continuations] Make use of `llvm::zip` in `CleanupContinuationsPass::updateCpsFunctionArgs` Adjust the llvm-tblgen memory limit lgc: split lowering of GroupMemcpyOp for mesh/task shaders to MeshTaskShader PatchResourceCollect: stop cleaning undef output value in some cases [Continuations] Freeze poison that is stored to payload Fix wrong pipeline dump message Fix issues with coherent Update SPIR-V header to latest version Simplify load of attribute ring buffer descriptor [Continuations] Enable Traversal specialization test Adjust tests to take into account upstream overload of rsrc and samp arguments [Continuations] Fix signed/unsigned comparison warning LowerRaytracingPipeline: tease apart return handling and any-hit exit handling LowerRaytracingPipeline: unify function end handling lgc: support int4 in cooperative matrix Add rounding mode for PackHalf2x16 [Continuations] Fix unused variable warning lgc: Add MsgPackScanner [Continuations] Remove `LegacyCleanupContinuations` pass LowerCooperativeMatrix: fix compile warning due to missing break statement lgc: New RegStackUsage to propagate reg/stack usage [LGC] Add helper lane state for subgroups Add client name to SPIR-V frontend Use more flexible readfirstlane Gate argument definition for `GEP::collectOffset`. [Continuations] Introduce dummy `csp` argument for `lgc.cps.jump` [Continuations] Tolerate non-waiting AwaitTraversal Add AmdExtD3DShaderIntrinsics_LoadDwordAtAddrx3 Add more BuiltIn which has primitive index in mesh shader Rewrite the helper getShaderModuleUsageInfo [Continuations] Replace isLgcRtOp helper with isDialectOp helper from dialects Downgrade scope Device to Workgroup if permitted Optimize performance for PrimSetup Update submodule llvm-dialects Remove inactive built-ins for last vertex processing stage Fix internal hash for color export shader Fix call to lookupIntrinsicID [Continuations] Move Simplifying GEP helpers to CompilerUtils [PatchBufferOp] Generate struct buffer cmpxchg intrinsics [LGC] Refactor default wave size setting [Continuations] Derive `DispatchSystemData` type from `_cont_DispatchRaysIndex3` [Continuations] Replace "lgc.rt" starts_with check Rename some classes and files Optimize PointSize write when the value is 1.0 [Continuations] Handle `_AmdGetShaderRecordIndex` calls [LowerBufferOperations] Check for uniform buffer pointers for s_buffer_load lgc: Mark applicable LgcDialect and Builder ops as NoDivergenceSource
qiaojbao
changed the title
Test summary for commit b86d36bCTS tests (Failed: 2/148731)
Ubuntu navi3x, SrdcvkUbuntu navi2x, Srdcvk |
Test summary for commit bc32393CTS tests (Failed: 2/148731)
Ubuntu navi3x, SrdcvkUbuntu navi2x, Srdcvk |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Set SLC=0 for ATM
LowerGpuRt: fix a type confusion
Honor NonUniform decorations on OpAccessChain operands Force WGP when NGG is in passthrough mode
Use idxen modifier for (RW)StructuredBuffers
Compatible with LLVM upstream change that moved Utility funcs to Utils Include Line Correlation Info for Tools via LLPC
[Continuations] Cleanup Complete op lowering
[CompilerUtils] Improve handling of freeze in ValueOriginTracker [compilerutils] Add GetGlobalInModule to CrossModuleInliner amdllpc: some usability improvements
[Continuations] Add SpecializeDriverShadersPass
Split up gl_in array type.
[Continuations] Remove stack lowering define
Use inbounds gep to index row major matrix
llvmraytracing: Remove support for _AmdContinuationsGetFlags Properly encapsulate state that affects GPURT specialization llvmraytracing: only support a waitmask of -1
llvmraytracing: remove remaining traces of EnqueueCall compilerutils: fix a warning
[LowerGlobals] Mark globals with
buffer.indexusers as readonly Implement structural GEP dialect for in/outllvmraytracing: add LLPC/LLVM scratch/global address spaces llpc/ProcessGpuRtLibrary: use earlyGpurtTransform
[Continuations] Make use of
llvm::zipinCleanupContinuationsPass::updateCpsFunctionArgsAdjust the llvm-tblgen memory limitlgc: split lowering of GroupMemcpyOp for mesh/task shaders to MeshTaskShader PatchResourceCollect: stop cleaning undef output value in some cases [Continuations] Freeze poison that is stored to payload Fix wrong pipeline dump message
Fix issues with coherent
Update SPIR-V header to latest version
Simplify load of attribute ring buffer descriptor
[Continuations] Enable Traversal specialization test Adjust tests to take into account upstream overload of rsrc and samp arguments [Continuations] Fix signed/unsigned comparison warning LowerRaytracingPipeline: tease apart return handling and any-hit exit handling LowerRaytracingPipeline: unify function end handling lgc: support int4 in cooperative matrix
Add rounding mode for PackHalf2x16
[Continuations] Fix unused variable warning
lgc: Add MsgPackScanner
[Continuations] Remove
LegacyCleanupContinuationspass LowerCooperativeMatrix: fix compile warning due to missing break statement lgc: New RegStackUsage to propagate reg/stack usage [LGC] Add helper lane state for subgroupsAdd client name to SPIR-V frontend
Use more flexible readfirstlane
Gate argument definition for
GEP::collectOffset. [Continuations] Introduce dummycspargument forlgc.cps.jump[Continuations] Tolerate non-waiting AwaitTraversal Add AmdExtD3DShaderIntrinsics_LoadDwordAtAddrx3Add more BuiltIn which has primitive index in mesh shader Rewrite the helper getShaderModuleUsageInfo
[Continuations] Replace isLgcRtOp helper with isDialectOp helper from dialects Downgrade scope Device to Workgroup if permitted
Optimize performance for PrimSetup
Update submodule llvm-dialects
Remove inactive built-ins for last vertex processing stage Fix internal hash for color export shader
Fix call to lookupIntrinsicID
[Continuations] Move Simplifying GEP helpers to CompilerUtils [PatchBufferOp] Generate struct buffer cmpxchg intrinsics [LGC] Refactor default wave size setting
[Continuations] Derive
DispatchSystemDatatype from_cont_DispatchRaysIndex3[Continuations] Replace "lgc.rt" starts_with check Rename some classes and filesOptimize PointSize write when the value is 1.0
[Continuations] Handle
_AmdGetShaderRecordIndexcalls [LowerBufferOperations] Check for uniform buffer pointers for s_buffer_load lgc: Mark applicable LgcDialect and Builder ops as NoDivergenceSource