This is an automated email from the git hooks/post-receive script.
unknown user pushed a change to branch master in repository llvm.
from c16c6e66658 [ValueTracking] Make unit tests easier to write; NFC new 83bcdb16b44 [InstSimplify] add tests for undef + partial undef constant [...] new 3ccf8b50bf5 [X86] Split skylake-avx512 run lines in SLP vectorizer test [...] new b8f365e6e8b [x86] add tests for undef + partial undef constant folding; NFC new 39a2bbdea0f [DA] GPUDivergenceAnalysis for unstructured GPU kernels new 37b386de213 AMDGPU: Fix various issues around the VirtReg2Value mapping new e3924b1c156 AMDGPU: Divergence-driven selection of scalar buffer load i [...]
The 6 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "adds" were already present in the repository and have only been added to this reference.
Summary of changes: include/llvm/Analysis/DivergenceAnalysis.h | 27 ++ include/llvm/Analysis/LegacyDivergenceAnalysis.h | 10 +- include/llvm/CodeGen/FunctionLoweringInfo.h | 1 + lib/Analysis/DivergenceAnalysis.cpp | 33 +++ lib/Analysis/LegacyDivergenceAnalysis.cpp | 101 +++++-- lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp | 13 +- lib/Target/AMDGPU/SIISelLowering.cpp | 172 +++++++----- lib/Target/AMDGPU/SIISelLowering.h | 4 +- lib/Target/AMDGPU/SIInstrInfo.cpp | 185 +------------ lib/Target/AMDGPU/SIInstrInfo.h | 2 - lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.cpp | 7 +- lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.h | 5 +- .../DivergenceAnalysis/AMDGPU/always_uniform.ll | 14 + test/Analysis/DivergenceAnalysis/AMDGPU/atomics.ll | 45 +++ .../DivergenceAnalysis/AMDGPU/hidden_diverge.ll | 26 ++ .../AMDGPU/hidden_loopdiverge.ll | 223 +++++++++++++++ .../DivergenceAnalysis/AMDGPU/intrinsics.ll | 13 + .../DivergenceAnalysis/AMDGPU/irreducible.ll | 48 ++++ .../DivergenceAnalysis/AMDGPU/kernel-args.ll | 41 +++ .../AMDGPU/lit.local.cfg | 0 .../AMDGPU/llvm.amdgcn.buffer.atomic.ll | 103 +++++++ .../AMDGPU/llvm.amdgcn.image.atomic.ll | 131 +++++++++ .../DivergenceAnalysis/AMDGPU/no-return-blocks.ll | 30 ++ .../DivergenceAnalysis/AMDGPU/phi-undef.ll | 31 +++ .../DivergenceAnalysis/AMDGPU/temporal_diverge.ll | 154 +++++++++++ .../AMDGPU/workitem-intrinsics.ll | 45 +++ test/Analysis/DivergenceAnalysis/NVPTX/daorder.ll | 47 ++++ test/Analysis/DivergenceAnalysis/NVPTX/diverge.ll | 175 ++++++++++++ .../DivergenceAnalysis/NVPTX/hidden_diverge.ll | 30 ++ .../DivergenceAnalysis/NVPTX/irreducible.ll | 55 ++++ .../NVPTX/lit.local.cfg | 0 test/CodeGen/AMDGPU/sdwa-peephole.ll | 7 +- test/CodeGen/AMDGPU/smrd-fold-offset.mir | 8 +- test/CodeGen/AMDGPU/smrd.ll | 45 +-- test/CodeGen/X86/fp-undef.ll | 90 ++++++ test/Transforms/InstSimplify/fp-undef.ll | 80 ++++++ test/Transforms/SLPVectorizer/X86/arith-add.ll | 3 +- test/Transforms/SLPVectorizer/X86/arith-mul.ll | 3 +- test/Transforms/SLPVectorizer/X86/arith-sub.ll | 3 +- test/Transforms/SLPVectorizer/X86/fabs.ll | 3 +- test/Transforms/SLPVectorizer/X86/fcopysign.ll | 3 +- test/Transforms/SLPVectorizer/X86/fma.ll | 3 +- test/Transforms/SLPVectorizer/X86/fptosi.ll | 147 +++++----- test/Transforms/SLPVectorizer/X86/fptoui.ll | 305 ++++++++++++--------- test/Transforms/SLPVectorizer/X86/fround.ll | 3 +- test/Transforms/SLPVectorizer/X86/shift-ashr.ll | 3 +- test/Transforms/SLPVectorizer/X86/shift-lshr.ll | 3 +- test/Transforms/SLPVectorizer/X86/shift-shl.ll | 3 +- test/Transforms/SLPVectorizer/X86/sitofp.ll | 234 +++++++++------- test/Transforms/SLPVectorizer/X86/uitofp.ll | 145 ++++++---- 50 files changed, 2195 insertions(+), 667 deletions(-) create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/always_uniform.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/atomics.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/hidden_diverge.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/hidden_loopdiverge.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/intrinsics.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/irreducible.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/kernel-args.ll copy test/Analysis/{CostModel => DivergenceAnalysis}/AMDGPU/lit.local.cfg (100%) create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/llvm.amdgcn.buffer.atomic.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/llvm.amdgcn.image.atomic.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/no-return-blocks.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/phi-undef.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/temporal_diverge.ll create mode 100644 test/Analysis/DivergenceAnalysis/AMDGPU/workitem-intrinsics.ll create mode 100644 test/Analysis/DivergenceAnalysis/NVPTX/daorder.ll create mode 100644 test/Analysis/DivergenceAnalysis/NVPTX/diverge.ll create mode 100644 test/Analysis/DivergenceAnalysis/NVPTX/hidden_diverge.ll create mode 100644 test/Analysis/DivergenceAnalysis/NVPTX/irreducible.ll copy test/Analysis/{LegacyDivergenceAnalysis => DivergenceAnalysis}/NVPTX/lit.loca [...]