[PATCH] D119235: [AMDGPU][NFC] Fix typos
Jay Foad via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Feb 8 05:22:11 PST 2022
foad added inline comments.
================
Comment at: llvm/lib/Target/AMDGPU/MIMGInstructions.td:229
// This class used to use !foldl to memoize the AddrAsmNames list.
-// It turned out that that was much slower than using !filter.
class MIMGNSAHelper<int num_addrs> {
----------------
This ones seems like a legitimate use of "that that".
================
Comment at: llvm/lib/Target/AMDGPU/R600ClauseMergePass.cpp:10
/// \file
-/// R600EmitClauseMarker pass emits CFAlu instruction in a conservative maneer.
+/// R600EmitClauseMarker pass emits CFAlu instruction in a conservative maner.
/// This pass is merging consecutive CFAlus where applicable.
----------------
Should be "manner".
================
Comment at: llvm/lib/Target/AMDGPU/R600InstrInfo.cpp:853
// be predicated. Until we have proper support for instruction clauses in the
- // backend, we will mark KILL* instructions as unpredicable.
----------------
No, we're talking about predication so "unpredicable" is correct.
================
Comment at: llvm/lib/Target/AMDGPU/SIInstrInfo.cpp:2954
// Multiplied part is the constant: Use v_madmk_{f16, f32}.
- // We should only expect these to be on src0 due to canonicalizations.
if (Src0->isReg() && Src0->getReg() == Reg) {
----------------
This one doesn't seem particularly wrong.
================
Comment at: llvm/lib/Target/AMDGPU/SIMachineScheduler.cpp:94
+// The OpenCL doc says for example latency of 400 cycles for a global mem
+// access, which is hidden by 10 instructions if the wavefront count is 10.
----------------
There is no typo fix here?
================
Comment at: llvm/lib/Target/AMDGPU/SIPeepholeSDWA.cpp:1143
+ // was already destroyed). So if SDWAOperand is also a potential MI then do
+ // not apply it.
if (PotentialMatches.count(Operand->getParentInst()) == 0)
----------------
No typo fixes here?
================
Comment at: llvm/lib/Target/AMDGPU/SIWholeQuadMode.cpp:972
if (IsDemote) {
- // Demotes deactive quads with only helper lanes
+ // Demotes deactivate quads with only helper lanes
LiveMaskWQM = MRI->createVirtualRegister(TRI->getBoolRC());
----------------
Should probably be "inactive"? Or perhaps "deactivated"?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D119235/new/
https://reviews.llvm.org/D119235
More information about the llvm-commits
mailing list