[InstSimplify] Add constant folding support for ucmp/scmp intrinsics
#93730
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
llvmbot
added
llvm:analysis
|
@llvm/pr-subscribers-llvm-transforms Author: None (Poseydon42) ChangesThis PR adds support for folding calls to Full diff: https://github.com/llvm/llvm-project/pull/93730.diff 2 Files Affected:
diff --git a/llvm/lib/Analysis/ConstantFolding.cpp b/llvm/lib/Analysis/ConstantFolding.cpp
index 046a769453808..c11d3032020f4 100644
--- a/llvm/lib/Analysis/ConstantFolding.cpp
+++ b/llvm/lib/Analysis/ConstantFolding.cpp
@@ -1506,6 +1506,8 @@ bool llvm::canConstantFoldCallTo(const CallBase *Call, const Function *F) {
case Intrinsic::smin:
case Intrinsic::umax:
case Intrinsic::umin:
+ case Intrinsic::scmp:
+ case Intrinsic::ucmp:
case Intrinsic::sadd_with_overflow:
case Intrinsic::uadd_with_overflow:
case Intrinsic::ssub_with_overflow:
@@ -2764,6 +2766,32 @@ static Constant *ConstantFoldIntrinsicCall2(Intrinsic::ID IntrinsicID, Type *Ty,
? *C0
: *C1);
+ case Intrinsic::scmp:
+ case Intrinsic::ucmp:
+ if (isa<PoisonValue>(Operands[0]) || isa<PoisonValue>(Operands[1]))
+ return PoisonValue::get(Ty);
+
+ if (!C0 || !C1)
+ return UndefValue::get(Ty);
+
+ if (ICmpInst::compare(*C0, *C1,
+ IntrinsicID == Intrinsic::scmp
+ ? ICmpInst::ICMP_SLT
+ : ICmpInst::ICMP_ULT)) {
+ return ConstantInt::get(Ty, -1, true);
+ }
+ if (ICmpInst::compare(*C0, *C1, ICmpInst::ICMP_EQ)) {
+ return ConstantInt::get(Ty, 0);
+ }
+ if (ICmpInst::compare(*C0, *C1,
+ IntrinsicID == Intrinsic::scmp
+ ? ICmpInst::ICMP_SGT
+ : ICmpInst::ICMP_UGT)) {
+ return ConstantInt::get(Ty, 1);
+ }
+
+ assert(false && "Integer values must compare as equal, or one must be less than the other");
+
case Intrinsic::usub_with_overflow:
case Intrinsic::ssub_with_overflow:
// X - undef -> { 0, false }
diff --git a/llvm/test/Transforms/InstSimplify/uscmp.ll b/llvm/test/Transforms/InstSimplify/uscmp.ll
new file mode 100644
index 0000000000000..c4e9f7bd54fd7
--- /dev/null
+++ b/llvm/test/Transforms/InstSimplify/uscmp.ll
@@ -0,0 +1,41 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt < %s -passes=instsimplify -S | FileCheck %s
+define i8 @scmp_lt() {
+; CHECK-LABEL: define i8 @scmp_lt() {
+; CHECK-NEXT: ret i8 -1
+;
+ %1 = call i8 @llvm.scmp(i32 -7, i32 3)
+ ret i8 %1
+}
+
+define i2 @ucmp_eq() {
+; CHECK-LABEL: define i2 @ucmp_eq() {
+; CHECK-NEXT: ret i2 0
+;
+ %1 = call i2 @llvm.ucmp(i32 12, i32 12)
+ ret i2 %1
+}
+
+define i100 @ucmp_gt() {
+; CHECK-LABEL: define i100 @ucmp_gt() {
+; CHECK-NEXT: ret i100 1
+;
+ %1 = call i100 @llvm.ucmp(i32 7, i32 3)
+ ret i100 %1
+}
+
+define <4 x i8> @ucmp_lt_splat() {
+; CHECK-LABEL: define <4 x i8> @ucmp_lt_splat() {
+; CHECK-NEXT: ret <4 x i8> <i8 -1, i8 -1, i8 -1, i8 -1>
+;
+ %1 = call <4 x i8> @llvm.ucmp(<4 x i32> splat(i32 1), <4 x i32> splat(i32 3))
+ ret <4 x i8> %1
+}
+
+define <4 x i8> @scmp_nonsplat() {
+; CHECK-LABEL: define <4 x i8> @scmp_nonsplat() {
+; CHECK-NEXT: ret <4 x i8> <i8 1, i8 0, i8 1, i8 -1>
+;
+ %1 = call <4 x i8> @llvm.scmp(<4 x i32> <i32 0, i32 1, i32 2, i32 3>, <4 x i32> <i32 -1, i32 1, i32 -2, i32 4>)
+ ret <4 x i8> %1
+}
|
|
@llvm/pr-subscribers-llvm-analysis Author: None (Poseydon42) ChangesThis PR adds support for folding calls to Full diff: https://github.com/llvm/llvm-project/pull/93730.diff 2 Files Affected:
diff --git a/llvm/lib/Analysis/ConstantFolding.cpp b/llvm/lib/Analysis/ConstantFolding.cpp
index 046a769453808..c11d3032020f4 100644
--- a/llvm/lib/Analysis/ConstantFolding.cpp
+++ b/llvm/lib/Analysis/ConstantFolding.cpp
@@ -1506,6 +1506,8 @@ bool llvm::canConstantFoldCallTo(const CallBase *Call, const Function *F) {
case Intrinsic::smin:
case Intrinsic::umax:
case Intrinsic::umin:
+ case Intrinsic::scmp:
+ case Intrinsic::ucmp:
case Intrinsic::sadd_with_overflow:
case Intrinsic::uadd_with_overflow:
case Intrinsic::ssub_with_overflow:
@@ -2764,6 +2766,32 @@ static Constant *ConstantFoldIntrinsicCall2(Intrinsic::ID IntrinsicID, Type *Ty,
? *C0
: *C1);
+ case Intrinsic::scmp:
+ case Intrinsic::ucmp:
+ if (isa<PoisonValue>(Operands[0]) || isa<PoisonValue>(Operands[1]))
+ return PoisonValue::get(Ty);
+
+ if (!C0 || !C1)
+ return UndefValue::get(Ty);
+
+ if (ICmpInst::compare(*C0, *C1,
+ IntrinsicID == Intrinsic::scmp
+ ? ICmpInst::ICMP_SLT
+ : ICmpInst::ICMP_ULT)) {
+ return ConstantInt::get(Ty, -1, true);
+ }
+ if (ICmpInst::compare(*C0, *C1, ICmpInst::ICMP_EQ)) {
+ return ConstantInt::get(Ty, 0);
+ }
+ if (ICmpInst::compare(*C0, *C1,
+ IntrinsicID == Intrinsic::scmp
+ ? ICmpInst::ICMP_SGT
+ : ICmpInst::ICMP_UGT)) {
+ return ConstantInt::get(Ty, 1);
+ }
+
+ assert(false && "Integer values must compare as equal, or one must be less than the other");
+
case Intrinsic::usub_with_overflow:
case Intrinsic::ssub_with_overflow:
// X - undef -> { 0, false }
diff --git a/llvm/test/Transforms/InstSimplify/uscmp.ll b/llvm/test/Transforms/InstSimplify/uscmp.ll
new file mode 100644
index 0000000000000..c4e9f7bd54fd7
--- /dev/null
+++ b/llvm/test/Transforms/InstSimplify/uscmp.ll
@@ -0,0 +1,41 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt < %s -passes=instsimplify -S | FileCheck %s
+define i8 @scmp_lt() {
+; CHECK-LABEL: define i8 @scmp_lt() {
+; CHECK-NEXT: ret i8 -1
+;
+ %1 = call i8 @llvm.scmp(i32 -7, i32 3)
+ ret i8 %1
+}
+
+define i2 @ucmp_eq() {
+; CHECK-LABEL: define i2 @ucmp_eq() {
+; CHECK-NEXT: ret i2 0
+;
+ %1 = call i2 @llvm.ucmp(i32 12, i32 12)
+ ret i2 %1
+}
+
+define i100 @ucmp_gt() {
+; CHECK-LABEL: define i100 @ucmp_gt() {
+; CHECK-NEXT: ret i100 1
+;
+ %1 = call i100 @llvm.ucmp(i32 7, i32 3)
+ ret i100 %1
+}
+
+define <4 x i8> @ucmp_lt_splat() {
+; CHECK-LABEL: define <4 x i8> @ucmp_lt_splat() {
+; CHECK-NEXT: ret <4 x i8> <i8 -1, i8 -1, i8 -1, i8 -1>
+;
+ %1 = call <4 x i8> @llvm.ucmp(<4 x i32> splat(i32 1), <4 x i32> splat(i32 3))
+ ret <4 x i8> %1
+}
+
+define <4 x i8> @scmp_nonsplat() {
+; CHECK-LABEL: define <4 x i8> @scmp_nonsplat() {
+; CHECK-NEXT: ret <4 x i8> <i8 1, i8 0, i8 1, i8 -1>
+;
+ %1 = call <4 x i8> @llvm.scmp(<4 x i32> <i32 0, i32 1, i32 2, i32 3>, <4 x i32> <i32 -1, i32 1, i32 -2, i32 4>)
+ ret <4 x i8> %1
+}
|
Poseydon42
force-pushed
the
uscmp-constant-folding
branch
from
b7ef565 to
28f7512
Compare
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
tschuett
reviewed
May 29, 2024
tschuett
reviewed
May 29, 2024
nikic
reviewed
May 29, 2024
Poseydon42
force-pushed
the
uscmp-constant-folding
branch
from
28f7512 to
d926bc1
Compare
nikic
approved these changes
May 30, 2024
Poseydon42
force-pushed
the
uscmp-constant-folding
branch
from
d926bc1 to
0da0092
Compare
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 adds support for folding calls to
ucmp/scmpintrinsics with constant arguments.