LLVM / project / a216358 / AMDGPU: Replace amdgpu-no-agpr with amdgpu-agpr-alloc (#129893)

LLVM/project a216358 — llvm/docs AMDGPUUsage.rst, llvm/lib/Target/AMDGPU AMDGPUAttributor.cpp SIRegisterInfo.cpp

10 days ago by Matt Arsenault via GitHub on ⎇

main

AMDGPU: Replace amdgpu-no-agpr with amdgpu-agpr-alloc (#129893)

This performs the minimal replacment of amdgpu-no-agpr to
amdgpu-agpr-alloc=0. Most of the test diffs are due to the new
attribute sorting later alphabetically.

We could do better by trying to perform range merging in the attributor,
and trying to pick non-0 values.

Delta		File
+21	-21	llvm/test/CodeGen/AMDGPU/propagate-waves-per-eu.ll
+21	-21	llvm/test/CodeGen/AMDGPU/annotate-kernel-features-hsa-call.ll
+13	-13	llvm/test/CodeGen/AMDGPU/implicitarg-offset-attributes.ll
+13	-13	llvm/test/CodeGen/AMDGPU/annotate-kernel-features-hsa.ll
+9	-9	llvm/test/CodeGen/AMDGPU/propagate-flat-work-group-size.ll
+9	-9	llvm/test/CodeGen/AMDGPU/annotate-kernel-features.ll
+7	-6	llvm/test/CodeGen/AMDGPU/amdgpu-no-agprs-violations.ll
+6	-6	llvm/test/CodeGen/AMDGPU/vgpr-agpr-limit-gfx90a.ll
+6	-6	llvm/test/CodeGen/AMDGPU/attributor-flatscratchinit.ll
+6	-6	llvm/test/CodeGen/AMDGPU/amdgpu-num-agpr.ll
+5	-5	llvm/test/CodeGen/AMDGPU/remove-no-kernel-id-attribute.ll
+7	-2	llvm/lib/Target/AMDGPU/AMDGPUAttributor.cpp
+1	-7	llvm/lib/Target/AMDGPU/SIRegisterInfo.cpp
+1	-6	llvm/docs/AMDGPUUsage.rst
+3	-3	llvm/test/CodeGen/AMDGPU/uniform-work-group-multistep.ll
+3	-3	llvm/test/CodeGen/AMDGPU/amdgpu-attributor-no-agpr.ll
+3	-3	llvm/test/CodeGen/AMDGPU/uniform-work-group-recursion-test.ll
+3	-3	llvm/test/CodeGen/AMDGPU/simple-indirect-call-2.ll
+4	-1	llvm/lib/Target/AMDGPU/SIMachineFunctionInfo.cpp
+2	-2	llvm/test/CodeGen/AMDGPU/agpr-copy-no-free-registers.ll
+2	-2	llvm/test/CodeGen/AMDGPU/uniform-work-group-propagate-attribute.ll
+2	-2	llvm/test/CodeGen/AMDGPU/uniform-work-group-prevent-attribute-propagation.ll
+2	-2	llvm/test/CodeGen/AMDGPU/issue120256-annotate-constexpr-addrspacecast.ll
+2	-2	llvm/test/CodeGen/AMDGPU/uniform-work-group-nested-function-calls.ll
+2	-2	llvm/test/CodeGen/AMDGPU/mfma-cd-select.ll
+2	-2	llvm/test/CodeGen/AMDGPU/addrspacecast-constantexpr.ll
+1	-1	llvm/test/CodeGen/AMDGPU/preload-implicit-kernargs.ll
+1	-1	llvm/test/CodeGen/AMDGPU/agpr-copy-no-free-registers-assertion-after-ra-failure.ll
+1	-1	llvm/test/CodeGen/AMDGPU/amdhsa-kernarg-preload-num-sgprs.ll
+1	-1	llvm/test/CodeGen/AMDGPU/captured-frame-index.ll
+1	-1	llvm/test/CodeGen/AMDGPU/copy-vgpr-clobber-spill-vgpr.mir
+1	-1	llvm/test/CodeGen/AMDGPU/direct-indirect-call.ll
+1	-1	llvm/test/CodeGen/AMDGPU/duplicate-attribute-indirect.ll
+1	-1	llvm/test/CodeGen/AMDGPU/indirect-call-set-from-other-function.ll
+1	-1	llvm/test/CodeGen/AMDGPU/invalid-hidden-kernarg-in-kernel-signature.ll
+1	-1	llvm/test/CodeGen/AMDGPU/mfma-bf16-vgpr-cd-select.ll
+1	-1	llvm/test/CodeGen/AMDGPU/mfma-vgpr-cd-select-gfx942.ll
+1	-1	llvm/test/CodeGen/AMDGPU/mfma-vgpr-cd-select.ll
+1	-1	llvm/test/CodeGen/AMDGPU/preload-kernargs.ll
+1	-1	llvm/test/CodeGen/AMDGPU/recursive_global_initializer.ll
+1	-1	llvm/test/CodeGen/AMDGPU/simple-indirect-call.ll
+1	-1	llvm/test/CodeGen/AMDGPU/smfmac_no_agprs.ll
+1	-1	llvm/test/CodeGen/AMDGPU/spill-regpressure-less.mir
+1	-1	llvm/test/CodeGen/AMDGPU/uniform-work-group-attribute-missing.ll
+1	-1	llvm/test/CodeGen/AMDGPU/uniform-work-group-test.ll
+174	-176	45 files

Unified Split Raw