This is an automated email from the git hooks/post-receive script.
unknown user pushed a change to branch master in repository llvm.
from a7bc052caeb gn build: Merge r357383 new fcc236c2688 [AMDGPU] Pre-allocate WWM registers to reduce VGPR pressure.
The 1 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: lib/Target/AMDGPU/AMDGPU.h | 6 +- lib/Target/AMDGPU/AMDGPUTargetMachine.cpp | 12 +- lib/Target/AMDGPU/CMakeLists.txt | 2 +- lib/Target/AMDGPU/SIFixWWMLiveness.cpp | 417 --------------------- lib/Target/AMDGPU/SIInstrInfo.cpp | 17 +- lib/Target/AMDGPU/SIInstructions.td | 7 + lib/Target/AMDGPU/SIMachineFunctionInfo.h | 5 + lib/Target/AMDGPU/SIPreAllocateWWMRegs.cpp | 221 +++++++++++ lib/Target/AMDGPU/SIRegisterInfo.cpp | 4 + lib/Target/AMDGPU/SIWholeQuadMode.cpp | 21 +- test/CodeGen/AMDGPU/atomic_optimizations_buffer.ll | 5 +- .../AMDGPU/atomic_optimizations_global_pointer.ll | 8 +- .../AMDGPU/atomic_optimizations_local_pointer.ll | 4 +- .../AMDGPU/atomic_optimizations_raw_buffer.ll | 4 +- .../AMDGPU/atomic_optimizations_struct_buffer.ll | 4 +- test/CodeGen/AMDGPU/fix-wwm-liveness.mir | 185 --------- test/CodeGen/AMDGPU/indirect-addressing-term.ll | 3 - test/CodeGen/AMDGPU/wqm.mir | 2 +- test/CodeGen/AMDGPU/wwm-reserved.ll | 188 ++++++++++ 19 files changed, 475 insertions(+), 640 deletions(-) delete mode 100644 lib/Target/AMDGPU/SIFixWWMLiveness.cpp create mode 100644 lib/Target/AMDGPU/SIPreAllocateWWMRegs.cpp delete mode 100644 test/CodeGen/AMDGPU/fix-wwm-liveness.mir create mode 100644 test/CodeGen/AMDGPU/wwm-reserved.ll