[PATCH] D93669: [AMDGPU] Support unaligned flat scratch in TLI
Stanislav Mekhanoshin via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Dec 22 10:34:50 PST 2020
rampitec marked an inline comment as done.
rampitec added inline comments.
================
Comment at: llvm/lib/Target/AMDGPU/SIISelLowering.cpp:1479
+ return AlignedBy4 ||
+ Subtarget->enableFlatScratch() ||
+ Subtarget->hasUnalignedScratchAccess();
----------------
arsenm wrote:
> Does it really only work if accessed through the flat instructions? I thought it would even through MUBUF
Yes, it is only flat scratch. For MUBUF spec says: "Swizzled addressing also requires dword aligned accesses".
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D93669/new/
https://reviews.llvm.org/D93669
More information about the llvm-commits
mailing list