[InstCombine] Do not fold logical is_finite test #136851
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
This was referenced Apr 23, 2025
|
No real-world impact. |
llvmbot
added
llvm:instcombine
|
@llvm/pr-subscribers-llvm-transforms Author: Yingwei Zheng (dtcxzyw) ChangesThis patch disables the fold for logical is_finite test (i.e., Alive2: https://alive2.llvm.org/ce/z/aH4LC7 Full diff: https://github.com/llvm/llvm-project/pull/136851.diff 2 Files Affected:
diff --git a/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp b/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
index f1b225c0f238a..979a9cbedf2ef 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineAndOrXor.cpp
@@ -1480,7 +1480,9 @@ Value *InstCombinerImpl::foldLogicOfFCmps(FCmpInst *LHS, FCmpInst *RHS,
}
}
- if (IsAnd && stripSignOnlyFPOps(LHS0) == stripSignOnlyFPOps(RHS0)) {
+ // This transform is not valid for a logical select.
+ if (!IsLogicalSelect && IsAnd &&
+ stripSignOnlyFPOps(LHS0) == stripSignOnlyFPOps(RHS0)) {
// and (fcmp ord x, 0), (fcmp u* x, inf) -> fcmp o* x, inf
// and (fcmp ord x, 0), (fcmp u* fabs(x), inf) -> fcmp o* x, inf
if (Value *Left = matchIsFiniteTest(Builder, LHS, RHS))
diff --git a/llvm/test/Transforms/InstCombine/and-fcmp.ll b/llvm/test/Transforms/InstCombine/and-fcmp.ll
index c7bbc8ab56f9a..ec1b6ad2ea168 100644
--- a/llvm/test/Transforms/InstCombine/and-fcmp.ll
+++ b/llvm/test/Transforms/InstCombine/and-fcmp.ll
@@ -4990,6 +4990,34 @@ define i1 @clang_builtin_isnormal_inf_check_copysign(half %x, half %y) {
ret i1 %and
}
+define i1 @clang_builtin_isnormal_inf_check_copysign_logical_select(half %x, half %y) {
+; CHECK-LABEL: @clang_builtin_isnormal_inf_check_copysign_logical_select(
+; CHECK-NEXT: [[COPYSIGN_X:%.*]] = call half @llvm.copysign.f16(half [[X:%.*]], half [[Y:%.*]])
+; CHECK-NEXT: [[ORD:%.*]] = fcmp ord half [[X]], 0xH0000
+; CHECK-NEXT: [[CMP:%.*]] = fcmp ueq half [[COPYSIGN_X]], 0xH7C00
+; CHECK-NEXT: [[AND:%.*]] = select i1 [[ORD]], i1 [[CMP]], i1 false
+; CHECK-NEXT: ret i1 [[AND]]
+;
+ %copysign.x = call half @llvm.copysign.f16(half %x, half %y)
+ %ord = fcmp ord half %x, 0.0
+ %cmp = fcmp uge half %copysign.x, 0xH7C00
+ %and = select i1 %ord, i1 %cmp, i1 false
+ ret i1 %and
+}
+
+define i1 @clang_builtin_isnormal_inf_check_fabs_nnan_logical_select(half %x) {
+; CHECK-LABEL: @clang_builtin_isnormal_inf_check_fabs_nnan_logical_select(
+; CHECK-NEXT: [[COPYSIGN_X:%.*]] = call half @llvm.fabs.f16(half [[X:%.*]])
+; CHECK-NEXT: [[AND:%.*]] = fcmp oeq half [[COPYSIGN_X]], 0xH7C00
+; CHECK-NEXT: ret i1 [[AND]]
+;
+ %copysign.x = call nnan half @llvm.fabs.f16(half %x)
+ %ord = fcmp ord half %x, 0.0
+ %cmp = fcmp uge half %copysign.x, 0xH7C00
+ %and = select i1 %ord, i1 %cmp, i1 false
+ ret i1 %and
+}
+
define i1 @isnormal_logical_select_0(half %x) {
; CHECK-LABEL: @isnormal_logical_select_0(
; CHECK-NEXT: [[FABS_X:%.*]] = call half @llvm.fabs.f16(half [[X:%.*]])
|
arsenm
approved these changes
Apr 23, 2025
nikic
approved these changes
Apr 23, 2025
|
/cherry-pick 8abc917 |
|
/pull-request #137606 |
swift-ci
pushed a commit
to swiftlang/llvm-project
that referenced
this pull request
Apr 29, 2025
This patch disables the fold for logical is_finite test (i.e., `and (fcmp ord x, 0), (fcmp u* x, inf) -> fcmp o* x, inf`). It is still possible to allow this fold for several logical cases (e.g., `stripSignOnlyFPOps(RHS0)` does not strip any operations). Since this patch has no real-world impact, I decided to disable this fold for all logical cases. Alive2: https://alive2.llvm.org/ce/z/aH4LC7 Closes llvm#136650. (cherry picked from commit 8abc917)
IanWood1
pushed a commit
to IanWood1/llvm-project
that referenced
this pull request
May 6, 2025
This patch disables the fold for logical is_finite test (i.e., `and (fcmp ord x, 0), (fcmp u* x, inf) -> fcmp o* x, inf`). It is still possible to allow this fold for several logical cases (e.g., `stripSignOnlyFPOps(RHS0)` does not strip any operations). Since this patch has no real-world impact, I decided to disable this fold for all logical cases. Alive2: https://alive2.llvm.org/ce/z/aH4LC7 Closes llvm#136650.
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 patch disables the fold for logical is_finite test (i.e.,
and (fcmp ord x, 0), (fcmp u* x, inf) -> fcmp o* x, inf).It is still possible to allow this fold for several logical cases (e.g.,
stripSignOnlyFPOps(RHS0)does not strip any operations). Since this patch has no real-world impact, I decided to disable this fold for all logical cases.Alive2: https://alive2.llvm.org/ce/z/aH4LC7
Closes #136650.