[InstCombine] Infer zext nneg flag directly #71906
Closed
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-llvm-transforms Author: Yingwei Zheng (dtcxzyw) ChangesThis PR sets the Alive2: https://alive2.llvm.org/ce/z/voT6HG Compile-time impact: https://llvm-compile-time-tracker.com/compare.php?from=dc3faf0ed0e3f1ea9e435a006167d9649f865da1&to=04c38cca49af96b6a66fc006576d8482d1725648&stat=instructions:u Full diff: https://github.com/llvm/llvm-project/pull/71906.diff 10 Files Affected:
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
index b0d31a92464e375..bde39606886ba0a 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
@@ -543,6 +543,8 @@ static Instruction *foldCttzCtlz(IntrinsicInst &II, InstCombinerImpl &IC) {
auto *Cttz = IC.Builder.CreateBinaryIntrinsic(Intrinsic::cttz, X,
IC.Builder.getTrue());
auto *ZextCttz = IC.Builder.CreateZExt(Cttz, II.getType());
+ if (auto *ZextInst = dyn_cast<PossiblyNonNegInst>(ZextCttz))
+ ZextInst->setNonNeg();
return IC.replaceInstUsesWith(II, ZextCttz);
}
@@ -639,7 +641,9 @@ static Instruction *foldCtpop(IntrinsicInst &II, InstCombinerImpl &IC) {
// ctpop (zext X) --> zext (ctpop X)
if (match(Op0, m_OneUse(m_ZExt(m_Value(X))))) {
Value *NarrowPop = IC.Builder.CreateUnaryIntrinsic(Intrinsic::ctpop, X);
- return CastInst::Create(Instruction::ZExt, NarrowPop, Ty);
+ Instruction *Zext = CastInst::Create(Instruction::ZExt, NarrowPop, Ty);
+ Zext->setNonNeg();
+ return Zext;
}
KnownBits Known(BitWidth);
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
index c59cf5a2a86ef44..a52b3fcb35234ae 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCasts.cpp
@@ -1159,7 +1159,9 @@ Instruction *InstCombinerImpl::visitZExt(ZExtInst &Zext) {
APInt AndValue(APInt::getLowBitsSet(SrcSize, MidSize));
Constant *AndConst = ConstantInt::get(A->getType(), AndValue);
Value *And = Builder.CreateAnd(A, AndConst, CSrc->getName() + ".mask");
- return new ZExtInst(And, DestTy);
+ auto *ZExt = new ZExtInst(And, DestTy);
+ ZExt->setNonNeg();
+ return ZExt;
}
if (SrcSize == DstSize) {
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
index 7c2ad92f919a3cc..bf69e467f9b6fed 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
@@ -5914,6 +5914,9 @@ static Instruction *processUMulZExtIdiom(ICmpInst &I, Value *MulVal,
APInt ShortMask = CI->getValue().trunc(MulWidth);
Value *ShortAnd = Builder.CreateAnd(Mul, ShortMask);
Value *Zext = Builder.CreateZExt(ShortAnd, BO->getType());
+ if (ShortMask.isNonNegative())
+ if (auto *ZextInst = dyn_cast<PossiblyNonNegInst>(Zext))
+ ZextInst->setNonNeg();
IC.replaceInstUsesWith(*BO, Zext);
} else {
llvm_unreachable("Unexpected Binary operation");
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineMulDivRem.cpp b/llvm/lib/Transforms/InstCombine/InstCombineMulDivRem.cpp
index db0804380855e3a..b471b2c887bd160 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineMulDivRem.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineMulDivRem.cpp
@@ -1243,7 +1243,12 @@ static Value *takeLog2(IRBuilderBase &Builder, Value *Op, unsigned Depth,
Value *X, *Y;
if (match(Op, m_ZExt(m_Value(X))))
if (Value *LogX = takeLog2(Builder, X, Depth, AssumeNonZero, DoFold))
- return IfFold([&]() { return Builder.CreateZExt(LogX, Op->getType()); });
+ return IfFold([&]() {
+ auto *V = Builder.CreateZExt(LogX, Op->getType());
+ if (auto *ZextInst = dyn_cast<PossiblyNonNegInst>(V))
+ ZextInst->setNonNeg();
+ return V;
+ });
// log2(X << Y) -> log2(X) + Y
// FIXME: Require one use unless X is 1?
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineShifts.cpp b/llvm/lib/Transforms/InstCombine/InstCombineShifts.cpp
index e178f9536b69f21..ad3ce748dcd96c1 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineShifts.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineShifts.cpp
@@ -409,6 +409,8 @@ Instruction *InstCombinerImpl::commonShiftTransforms(BinaryOperator &I) {
Value *Y;
if (match(Op1, m_OneUse(m_SExt(m_Value(Y))))) {
Value *NewExt = Builder.CreateZExt(Y, Ty, Op1->getName());
+ if (auto *ZextInst = dyn_cast<PossiblyNonNegInst>(NewExt))
+ ZextInst->setNonNeg();
return BinaryOperator::Create(I.getOpcode(), Op0, NewExt);
}
@@ -1331,7 +1333,10 @@ Instruction *InstCombinerImpl::visitLShr(BinaryOperator &I) {
"Big shift not simplified to zero?");
// lshr (zext iM X to iN), C --> zext (lshr X, C) to iN
Value *NewLShr = Builder.CreateLShr(X, ShAmtC);
- return new ZExtInst(NewLShr, Ty);
+ auto *Zext = new ZExtInst(NewLShr, Ty);
+ if (ShAmtC)
+ Zext->setNonNeg();
+ return Zext;
}
if (match(Op0, m_SExt(m_Value(X)))) {
@@ -1349,7 +1354,9 @@ Instruction *InstCombinerImpl::visitLShr(BinaryOperator &I) {
// zeros? lshr (sext iM X to iN), N-1 --> zext (lshr X, M-1) to iN
if (ShAmtC == BitWidth - 1) {
Value *NewLShr = Builder.CreateLShr(X, SrcTyBitWidth - 1);
- return new ZExtInst(NewLShr, Ty);
+ auto *Zext = new ZExtInst(NewLShr, Ty);
+ Zext->setNonNeg();
+ return Zext;
}
// lshr (sext iM X to iN), N-M --> zext (ashr X, min(N-M, M-1)) to iN
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp b/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
index 29cf04d82b2e401..248e3f8f7075cf3 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
@@ -481,6 +481,8 @@ Value *InstCombinerImpl::SimplifyDemandedUseBits(Value *V, APInt DemandedMask,
DemandedMask.getActiveBits() <= SrcBitWidth) {
// Convert to ZExt cast.
CastInst *NewCast = new ZExtInst(I->getOperand(0), VTy, I->getName());
+ if (InputKnown.isNonNegative())
+ NewCast->setNonNeg();
return InsertNewInstWith(NewCast, I->getIterator());
}
diff --git a/llvm/test/Transforms/InstCombine/div-shift.ll b/llvm/test/Transforms/InstCombine/div-shift.ll
index d208837f04594ab..9610746811a43a6 100644
--- a/llvm/test/Transforms/InstCombine/div-shift.ll
+++ b/llvm/test/Transforms/InstCombine/div-shift.ll
@@ -38,7 +38,7 @@ define <2 x i32> @t1vec(<2 x i16> %x, <2 x i32> %y) {
; rdar://11721329
define i64 @t2(i64 %x, i32 %y) {
; CHECK-LABEL: @t2(
-; CHECK-NEXT: [[TMP1:%.*]] = zext i32 [[Y:%.*]] to i64
+; CHECK-NEXT: [[TMP1:%.*]] = zext nneg i32 [[Y:%.*]] to i64
; CHECK-NEXT: [[TMP2:%.*]] = lshr i64 [[X:%.*]], [[TMP1]]
; CHECK-NEXT: ret i64 [[TMP2]]
;
@@ -52,7 +52,7 @@ define i64 @t2(i64 %x, i32 %y) {
define i64 @t3(i64 %x, i32 %y) {
; CHECK-LABEL: @t3(
; CHECK-NEXT: [[TMP1:%.*]] = add i32 [[Y:%.*]], 2
-; CHECK-NEXT: [[TMP2:%.*]] = zext i32 [[TMP1]] to i64
+; CHECK-NEXT: [[TMP2:%.*]] = zext nneg i32 [[TMP1]] to i64
; CHECK-NEXT: [[TMP3:%.*]] = lshr i64 [[X:%.*]], [[TMP2]]
; CHECK-NEXT: ret i64 [[TMP3]]
;
diff --git a/llvm/test/Transforms/InstCombine/load-cmp.ll b/llvm/test/Transforms/InstCombine/load-cmp.ll
index 56f6e042b3cafdb..e941284a798ed11 100644
--- a/llvm/test/Transforms/InstCombine/load-cmp.ll
+++ b/llvm/test/Transforms/InstCombine/load-cmp.ll
@@ -122,7 +122,7 @@ define i1 @test4(i32 %X) {
define i1 @test4_i16(i16 %X) {
; CHECK-LABEL: @test4_i16(
-; CHECK-NEXT: [[TMP1:%.*]] = zext i16 [[X:%.*]] to i32
+; CHECK-NEXT: [[TMP1:%.*]] = zext nneg i16 [[X:%.*]] to i32
; CHECK-NEXT: [[TMP2:%.*]] = lshr i32 933, [[TMP1]]
; CHECK-NEXT: [[TMP3:%.*]] = and i32 [[TMP2]], 1
; CHECK-NEXT: [[R:%.*]] = icmp ne i32 [[TMP3]], 0
diff --git a/llvm/test/Transforms/InstCombine/shift-by-signext.ll b/llvm/test/Transforms/InstCombine/shift-by-signext.ll
index b72f33fc6502364..7fe4364cc08016d 100644
--- a/llvm/test/Transforms/InstCombine/shift-by-signext.ll
+++ b/llvm/test/Transforms/InstCombine/shift-by-signext.ll
@@ -6,7 +6,7 @@
define i32 @t0_shl(i32 %x, i8 %shamt) {
; CHECK-LABEL: @t0_shl(
-; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext i8 [[SHAMT:%.*]] to i32
+; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext nneg i8 [[SHAMT:%.*]] to i32
; CHECK-NEXT: [[R:%.*]] = shl i32 [[X:%.*]], [[SHAMT_WIDE1]]
; CHECK-NEXT: ret i32 [[R]]
;
@@ -16,7 +16,7 @@ define i32 @t0_shl(i32 %x, i8 %shamt) {
}
define i32 @t1_lshr(i32 %x, i8 %shamt) {
; CHECK-LABEL: @t1_lshr(
-; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext i8 [[SHAMT:%.*]] to i32
+; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext nneg i8 [[SHAMT:%.*]] to i32
; CHECK-NEXT: [[R:%.*]] = lshr i32 [[X:%.*]], [[SHAMT_WIDE1]]
; CHECK-NEXT: ret i32 [[R]]
;
@@ -26,7 +26,7 @@ define i32 @t1_lshr(i32 %x, i8 %shamt) {
}
define i32 @t2_ashr(i32 %x, i8 %shamt) {
; CHECK-LABEL: @t2_ashr(
-; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext i8 [[SHAMT:%.*]] to i32
+; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext nneg i8 [[SHAMT:%.*]] to i32
; CHECK-NEXT: [[R:%.*]] = ashr i32 [[X:%.*]], [[SHAMT_WIDE1]]
; CHECK-NEXT: ret i32 [[R]]
;
@@ -37,7 +37,7 @@ define i32 @t2_ashr(i32 %x, i8 %shamt) {
define <2 x i32> @t3_vec_shl(<2 x i32> %x, <2 x i8> %shamt) {
; CHECK-LABEL: @t3_vec_shl(
-; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext <2 x i8> [[SHAMT:%.*]] to <2 x i32>
+; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext nneg <2 x i8> [[SHAMT:%.*]] to <2 x i32>
; CHECK-NEXT: [[R:%.*]] = shl <2 x i32> [[X:%.*]], [[SHAMT_WIDE1]]
; CHECK-NEXT: ret <2 x i32> [[R]]
;
@@ -47,7 +47,7 @@ define <2 x i32> @t3_vec_shl(<2 x i32> %x, <2 x i8> %shamt) {
}
define <2 x i32> @t4_vec_lshr(<2 x i32> %x, <2 x i8> %shamt) {
; CHECK-LABEL: @t4_vec_lshr(
-; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext <2 x i8> [[SHAMT:%.*]] to <2 x i32>
+; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext nneg <2 x i8> [[SHAMT:%.*]] to <2 x i32>
; CHECK-NEXT: [[R:%.*]] = lshr <2 x i32> [[X:%.*]], [[SHAMT_WIDE1]]
; CHECK-NEXT: ret <2 x i32> [[R]]
;
@@ -57,7 +57,7 @@ define <2 x i32> @t4_vec_lshr(<2 x i32> %x, <2 x i8> %shamt) {
}
define <2 x i32> @t5_vec_ashr(<2 x i32> %x, <2 x i8> %shamt) {
; CHECK-LABEL: @t5_vec_ashr(
-; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext <2 x i8> [[SHAMT:%.*]] to <2 x i32>
+; CHECK-NEXT: [[SHAMT_WIDE1:%.*]] = zext nneg <2 x i8> [[SHAMT:%.*]] to <2 x i32>
; CHECK-NEXT: [[R:%.*]] = ashr <2 x i32> [[X:%.*]], [[SHAMT_WIDE1]]
; CHECK-NEXT: ret <2 x i32> [[R]]
;
diff --git a/llvm/test/Transforms/InstCombine/vector-udiv.ll b/llvm/test/Transforms/InstCombine/vector-udiv.ll
index c3468e95ba3d4fd..c817b3a1ac5a0a4 100644
--- a/llvm/test/Transforms/InstCombine/vector-udiv.ll
+++ b/llvm/test/Transforms/InstCombine/vector-udiv.ll
@@ -75,7 +75,7 @@ define <4 x i32> @test_v4i32_shl_const_pow2(<4 x i32> %a0, <4 x i32> %a1) {
define <4 x i32> @test_v4i32_zext_shl_splatconst_pow2(<4 x i32> %a0, <4 x i16> %a1) {
; CHECK-LABEL: @test_v4i32_zext_shl_splatconst_pow2(
; CHECK-NEXT: [[TMP1:%.*]] = add <4 x i16> [[A1:%.*]], <i16 2, i16 2, i16 2, i16 2>
-; CHECK-NEXT: [[TMP2:%.*]] = zext <4 x i16> [[TMP1]] to <4 x i32>
+; CHECK-NEXT: [[TMP2:%.*]] = zext nneg <4 x i16> [[TMP1]] to <4 x i32>
; CHECK-NEXT: [[TMP3:%.*]] = lshr <4 x i32> [[A0:%.*]], [[TMP2]]
; CHECK-NEXT: ret <4 x i32> [[TMP3]]
;
@@ -88,7 +88,7 @@ define <4 x i32> @test_v4i32_zext_shl_splatconst_pow2(<4 x i32> %a0, <4 x i16> %
define <4 x i32> @test_v4i32_zext_shl_const_pow2(<4 x i32> %a0, <4 x i16> %a1) {
; CHECK-LABEL: @test_v4i32_zext_shl_const_pow2(
; CHECK-NEXT: [[TMP1:%.*]] = add <4 x i16> [[A1:%.*]], <i16 2, i16 3, i16 4, i16 5>
-; CHECK-NEXT: [[TMP2:%.*]] = zext <4 x i16> [[TMP1]] to <4 x i32>
+; CHECK-NEXT: [[TMP2:%.*]] = zext nneg <4 x i16> [[TMP1]] to <4 x i32>
; CHECK-NEXT: [[TMP3:%.*]] = lshr <4 x i32> [[A0:%.*]], [[TMP2]]
; CHECK-NEXT: ret <4 x i32> [[TMP3]]
;
|
dtcxzyw
force-pushed
the
perf/infer-nneg
branch
from
04c38cc to
07958d9
Compare
nikic
approved these changes
Nov 10, 2023
dtcxzyw
added a commit
that referenced
this pull request
Nov 13, 2023
This patch sets `nneg` flag when the zext is only used by a shift. Alive2: https://alive2.llvm.org/ce/z/h3xKjP Compile-time impact: https://llvm-compile-time-tracker.com/compare.php?from=bd611264993f64decbce178d460caf1d1cb05f59&to=26bc473b239010bb24ff1bc39d58b42ecbbc4730&stat=instructions:u This is an alternative to #71906.
|
Closed as it failed to improve compile time. Test changes are handled by #71947. |
zahiraam
pushed a commit
to zahiraam/llvm-project
that referenced
this pull request
Nov 20, 2023
This patch sets `nneg` flag when the zext is only used by a shift. Alive2: https://alive2.llvm.org/ce/z/h3xKjP Compile-time impact: https://llvm-compile-time-tracker.com/compare.php?from=bd611264993f64decbce178d460caf1d1cb05f59&to=26bc473b239010bb24ff1bc39d58b42ecbbc4730&stat=instructions:u This is an alternative to llvm#71906.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This PR sets the
nnegflag directly to salvage information and avoid recomputingisKnownNonNegative, which was introduced by #71534.Alive2: https://alive2.llvm.org/ce/z/voT6HG
Compile-time impact: https://llvm-compile-time-tracker.com/compare.php?from=bd611264993f64decbce178d460caf1d1cb05f59&to=07958d9959d8af31f599acef96d580cfe786c677&stat=instructions:u