[PATCH] D101292: [AMDGPU] Restrict immediate scratch offsets
Sebastian Neubauer via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Mon Apr 26 06:42:41 PDT 2021
sebastian-ne created this revision.
sebastian-ne added reviewers: rampitec, arsenm, foad.
Herald added subscribers: kerbowa, hiraditya, t-tye, Anastasia, tpr, dstuttard, yaxunl, nhaehnle, jvesely, kzhuravl.
sebastian-ne requested review of this revision.
Herald added subscribers: llvm-commits, wdng.
Herald added a project: LLVM.
gfx9 does not work with negative offsets, gfx10 works only with
aligned negative offsets, but not with aligned negative offsets.
This is slightly more conservative than needed, gfx9 does support
negative offsets when a VGPR address is used and gfx10 supports
negative, unaligned offsets when an SGPR address is used.
I don't know if later gfx9.x cards still have these issues, I can provide an OpenCL test application if someone has the hardware.
Repository:
rG LLVM Github Monorepo
https://reviews.llvm.org/D101292
Files:
llvm/lib/Target/AMDGPU/AMDGPU.td
llvm/lib/Target/AMDGPU/AMDGPUISelDAGToDAG.cpp
llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
llvm/lib/Target/AMDGPU/GCNSubtarget.h
llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
llvm/test/CodeGen/AMDGPU/flat-scratch.ll
llvm/test/CodeGen/AMDGPU/local-stack-alloc-block-sp-reference.ll
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D101292.340504.patch
Type: text/x-patch
Size: 61577 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210426/09be7ebf/attachment-0001.bin>
More information about the llvm-commits
mailing list