[AMDGPU][NFC] Refactor SCC optimization #165871
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
@llvm/pr-subscribers-backend-amdgpu Author: None (LU-JOHN) ChangesRefactor SCC optimization Full diff: https://github.com/llvm/llvm-project/pull/165871.diff 1 Files Affected:
diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
index d930a21c2d7f5..9dd3bedf38bd7 100644
--- a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -10628,7 +10628,31 @@ bool SIInstrInfo::optimizeCompareInstr(MachineInstr &CmpInstr, Register SrcReg,
if (SrcReg2 && !getFoldableImm(SrcReg2, *MRI, CmpValue))
return false;
- const auto optimizeCmpSelect = [&CmpInstr, SrcReg, CmpValue, MRI,
+ // SCC is already valid after SCCValid.
+ // SCCRedefine will redefine SCC to the same value already available after
+ // SCCValid. If there are no intervening SCC conflicts delete SCCRedefine and
+ // update kill/dead flags if necessary.
+ const auto optimizeSCC = [this](MachineInstr *SCCValid,
+ MachineInstr *SCCRedefine) -> bool {
+ MachineInstr *KillsSCC = nullptr;
+ for (MachineInstr &MI : make_range(std::next(SCCValid->getIterator()),
+ SCCRedefine->getIterator())) {
+ if (MI.modifiesRegister(AMDGPU::SCC, &RI))
+ return false;
+ if (MI.killsRegister(AMDGPU::SCC, &RI))
+ KillsSCC = &MI;
+ }
+ if (MachineOperand *SccDef =
+ SCCValid->findRegisterDefOperand(AMDGPU::SCC, /*TRI=*/nullptr))
+ SccDef->setIsDead(false);
+ if (KillsSCC)
+ KillsSCC->clearRegisterKills(AMDGPU::SCC, /*TRI=*/nullptr);
+ SCCRedefine->eraseFromParent();
+
+ return true;
+ };
+
+ const auto optimizeCmpSelect = [&CmpInstr, SrcReg, CmpValue, MRI, optimizeSCC,
this]() -> bool {
if (CmpValue != 0)
return false;
@@ -10663,25 +10687,13 @@ bool SIInstrInfo::optimizeCompareInstr(MachineInstr &CmpInstr, Register SrcReg,
if (!setsSCCifResultIsNonZero(*Def) && !foldableSelect(Def))
return false;
- MachineInstr *KillsSCC = nullptr;
- for (MachineInstr &MI :
- make_range(std::next(Def->getIterator()), CmpInstr.getIterator())) {
- if (MI.modifiesRegister(AMDGPU::SCC, &RI))
- return false;
- if (MI.killsRegister(AMDGPU::SCC, &RI))
- KillsSCC = &MI;
- }
+ if (!optimizeSCC(Def, &CmpInstr))
+ return false;
- if (MachineOperand *SccDef =
- Def->findRegisterDefOperand(AMDGPU::SCC, /*TRI=*/nullptr))
- SccDef->setIsDead(false);
- if (KillsSCC)
- KillsSCC->clearRegisterKills(AMDGPU::SCC, /*TRI=*/nullptr);
- CmpInstr.eraseFromParent();
return true;
};
- const auto optimizeCmpAnd = [&CmpInstr, SrcReg, CmpValue, MRI,
+ const auto optimizeCmpAnd = [&CmpInstr, SrcReg, CmpValue, MRI, optimizeSCC,
this](int64_t ExpectedValue, unsigned SrcSize,
bool IsReversible, bool IsSigned) -> bool {
// s_cmp_eq_u32 (s_and_b32 $src, 1 << n), 1 << n => s_and_b32 $src, 1 << n
@@ -10755,21 +10767,8 @@ bool SIInstrInfo::optimizeCompareInstr(MachineInstr &CmpInstr, Register SrcReg,
if (IsReversedCC && !MRI->hasOneNonDBGUse(DefReg))
return false;
- MachineInstr *KillsSCC = nullptr;
- for (MachineInstr &MI :
- make_range(std::next(Def->getIterator()), CmpInstr.getIterator())) {
- if (MI.modifiesRegister(AMDGPU::SCC, &RI))
- return false;
- if (MI.killsRegister(AMDGPU::SCC, &RI))
- KillsSCC = &MI;
- }
-
- MachineOperand *SccDef =
- Def->findRegisterDefOperand(AMDGPU::SCC, /*TRI=*/nullptr);
- SccDef->setIsDead(false);
- if (KillsSCC)
- KillsSCC->clearRegisterKills(AMDGPU::SCC, /*TRI=*/nullptr);
- CmpInstr.eraseFromParent();
+ if (!optimizeSCC(Def, &CmpInstr))
+ return false;
if (!MRI->use_nodbg_empty(DefReg)) {
assert(!IsReversedCC);
|
|
@jmmartinez I don't think optimizeSCC will be used outside of optimizeCompareInstr, so I would prefer to keep it a lambda here. |
And to a static function only visible inside Having lambdas taking other lambdas by copy seems odd (specially since something simpler would do the job just as well). |
Made optimizeSCC and foldableSelect static functions. |
arsenm
approved these changes
Oct 31, 2025
Signed-off-by: John Lu <John.Lu@amd.com>
Signed-off-by: John Lu <John.Lu@amd.com>
Signed-off-by: John Lu <John.Lu@amd.com>
Signed-off-by: John Lu <John.Lu@amd.com>
Signed-off-by: John Lu <John.Lu@amd.com>
arsenm
approved these changes
Oct 31, 2025
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/138/builds/21197 Here is the relevant piece of the build log for the reference |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Refactor SCC optimization