[llvm] c1c49a3 - [AMDGPU] Fix comment type in the DSInstructions.td. NFC.
Stanislav Mekhanoshin via llvm-commits
llvm-commits at lists.llvm.org
Mon Apr 18 14:28:49 PDT 2022
Author: Stanislav Mekhanoshin
Date: 2022-04-18T14:28:12-07:00
New Revision: c1c49a356162b22554088d269f7689bdb044a9f1
URL: https://github.com/llvm/llvm-project/commit/c1c49a356162b22554088d269f7689bdb044a9f1
DIFF: https://github.com/llvm/llvm-project/commit/c1c49a356162b22554088d269f7689bdb044a9f1.diff
LOG: [AMDGPU] Fix comment type in the DSInstructions.td. NFC.
Added:
Modified:
llvm/lib/Target/AMDGPU/DSInstructions.td
Removed:
################################################################################
diff --git a/llvm/lib/Target/AMDGPU/DSInstructions.td b/llvm/lib/Target/AMDGPU/DSInstructions.td
index 1fb36991403ab..4e7a9b5a65cd8 100644
--- a/llvm/lib/Target/AMDGPU/DSInstructions.td
+++ b/llvm/lib/Target/AMDGPU/DSInstructions.td
@@ -885,9 +885,9 @@ defm : DSWritePat_mc <DS_WRITE_B96, vt, "store_local">;
}
// Select 128 bit loads and stores aligned less than 4 as a single ds_read_b128/
-// ds_write_b128 as a single instruction as this is faster than ds_read2_b64/
-// ds_write2_b64 which would be used otherwise. In this case a b64 access would
-// still be misaligned, but we will have 2 of them.
+// ds_write_b128 instruction as this is faster than ds_read2_b64/ds_write2_b64
+// which would be used otherwise. In this case a b64 access would still be
+// misaligned, but we will have 2 of them.
foreach vt = VReg_128.RegTypes in {
defm : DSReadPat_mc <DS_READ_B128, vt, "load_align_less_than_4_local">;
defm : DSWritePat_mc <DS_WRITE_B128, vt, "store_align_less_than_4_local">;
More information about the llvm-commits
mailing list