Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Promotion from AMD internal branch for 2024.Q2.2 #2901

Closed
wants to merge 1 commit into from
Closed

Commits on Jun 4, 2024

  1. Update llpc from commit e48a029e

    [Continuations] Add a test with boolean payload
    [Continuations] Add assertion on payload register count usage
    [Continuations] Add getUpperBoundOnTraceRayPayloadRegisters()
    [Continuations] Add shader jump scheduling workflow
    [Continuations] Add simple lit test for `_cont_ExitRayGen` inlining
    [Continuations] Add stable returnaddr argument
    [Continuations] Cleanup return in `else` block
    [Continuations] Cleanup use of `EntriesWithContinuationsFunctions`
    [Continuations] Dialectify continuation.return.
    [Continuations] Fix a regression caused by strip_whitelist.txt clean-up in GPURT
    [Continuations] Fix Cps Stack Lowering for KernelEntry
    [Continuations] Fix lit test for old LLVM
    [Continuations] Fix Traversal register count metadata
    [Continuations] Introduce `llvm::replaceCallsToFunction` helper
    [Continuations] Introduce `PayloadHelper`
    [Continuations] Make dialect stackstore take no pointer
    [Continuations] Move payload intrinsic handling to `LowerRaytracingPipeline`
    [Continuations] Move static helper function
    [Continuations] PAQ: Represent uniform structs by a single PAQ node
    [Continuations] Pass csp argument to Raygen, remove `_cont_SetupRayGen`
    [Continuations] Refactor `LowerAwait` pass
    [Continuations] Refactor PAQ node creation
    [Continuations] Remove unused includes
    [Continuations] Replace `forEachCall` with call to `replaceCallsToFunction`
    [Continuations] Return `PreservedAnalyes` directly.
    [Continuations] Run `LowerRaytracingPipeline` on Traversal
    [Continuations] Set preserved register count for IS
    [Continuations] Support _cont_ShaderStart
    [Continuations] Support GPURT options
    [Continuations] Use --lint-abort-on-error in lit tests
    [Continuations] Use as.continuation.reference consistently
    [Continuations] Use lgc.rt max payload size metadata
    [Continuations] Use max hit attributes from module
    [Continuations] Use updated update_test_checks.py to generate lit files
    Add SqBufRsrcWord3 initialization
    Add support for AmdExtD3DShaderIntrinsics_WaveScan Intrinsic
    Enable scalarization of descriptor loads
    Fix AmdTraceRayGetStaticId in ray query
    Fix invalid startswith call
    Fix static check error: Using 'memset' on struct that contains a 'std::string'
    Fix the issue where a maxVertOut value of zero leads to a division by maxVertOut, causing a crash
    Handle upstream llvm changes for metadata
    lgc: handle indirect load case in stridedIndexAdd
    lgc: Add BufferAddrToPtrOp and ExtendAddressOp
    lgc: Add BuilderCommon::CreateBuildVector
    lgc: Add debugging log for PatchBufferOp
    lgc: Add offset and ascii chars comment to bytes of data line
    lgc: emit amdgcn.fdot2.f16.f16 intrinsics
    lgc: Enable sinking in SimplifyCFG
    lgc: Extend lgc dialect with subgroupAll, AllEqual, Rotate
    lgc: Fix 'Add DbgInfo api to ModuleBunch'
    lgc: Fix crash in HLK Dot2AddHalfTest on GFX10.1
    lgc: Fix IO error if pre-rast shader don't export layer
    lgc: Fix the hang when shaderDBG is ON
    lgc: Move CreateIf() to BuilderCommon so a front-end can use it
    lgc: Refactor CreateSubgroupClustered* with contant cluster size
    lgc: Some coding style changes in PatchEntryPointMutate
    lgc: Use rtz version 16bit interpolation
    llvmraytracing: Allow setLgcRtShaderStage/getLgcRtShaderStage on GlobalObject
    Refactor for CooperativeMatrix using LGC_Dialect
    Remove the macro VKI_COOPERATIVE_MATRIX gating under lgc/llpc test cases
    Replace struct_buffer_load with s_buffer_load for strided buffer in some case
    Revert 'Pack mesh shader outputs in LDS space' to fix corruption in The Talos Principle 2
    Revert 'lgc: Add LoadBufferAddrOp'
    Rework SW stream-out
    Some coding changes for dynamic topology
    SPIRVReader: Map spirv device scope to agent
    Support dumping and reading GPURT options
    Support dynamic primitive topology when xfb is enabled
    Support uint64 AmdTraceRaySampleGpuTimer()
    Update tests affected by change for range attr in ctpop
    Update tests for upstream GEP change
    qiaojbao committed Jun 4, 2024
    Configuration menu
    Copy the full SHA
    6c770c7 View commit details
    Browse the repository at this point in the history