Skip to content

[Scalarizer] Make *_with_overflow intrinsics scalarizable #126815

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Feb 13, 2025
Merged

[Scalarizer] Make *_with_overflow intrinsics scalarizable #126815

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
1ee9014
Make uadd_with_overflow scalarizable
Icohedron Feb 11, 2025
e91e4c6
Add a test for an intrinsic that is not marked as isTriviallyScalariz…
Icohedron Feb 11, 2025
b2a128b
Make sadd, smul, ssub, umul, and usub trivially scalarizable
Icohedron Feb 12, 2025
0ecf77e
Remove unnecessary include from VectorUtils.cpp
Icohedron Feb 12, 2025
0d6eb29
Replace single quotes with double quotes so the update_test_checks.py…
Icohedron Feb 13, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions llvm/lib/Analysis/VectorUtils.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -125,6 +125,12 @@ bool llvm::isTriviallyScalarizable(Intrinsic::ID ID,
// https://github.com/llvm/llvm-project/issues/112408
switch (ID) {
case Intrinsic::frexp:
case Intrinsic::uadd_with_overflow:
case Intrinsic::sadd_with_overflow:
case Intrinsic::ssub_with_overflow:
case Intrinsic::usub_with_overflow:
case Intrinsic::umul_with_overflow:
case Intrinsic::smul_with_overflow:
return true;
}
return false;
Expand Down
24 changes: 24 additions & 0 deletions llvm/test/Transforms/Scalarizer/sadd_with_overflow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

define <3 x i32> @test_(<3 x i32> %a, <3 x i32> %b) {
; CHECK-LABEL: define <3 x i32> @test_(
; CHECK-SAME: <3 x i32> [[A:%.*]], <3 x i32> [[B:%.*]]) {
; CHECK-NEXT: [[B_I0:%.*]] = extractelement <3 x i32> [[B]], i64 0
; CHECK-NEXT: [[R_I0:%.*]] = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 [[B_I0]], i32 [[B_I0]])
; CHECK-NEXT: [[B_I1:%.*]] = extractelement <3 x i32> [[B]], i64 1
; CHECK-NEXT: [[R_I1:%.*]] = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 [[B_I1]], i32 [[B_I1]])
; CHECK-NEXT: [[B_I2:%.*]] = extractelement <3 x i32> [[B]], i64 2
; CHECK-NEXT: [[R_I2:%.*]] = call { i32, i1 } @llvm.sadd.with.overflow.i32(i32 [[B_I2]], i32 [[B_I2]])
; CHECK-NEXT: [[EL_ELEM0:%.*]] = extractvalue { i32, i1 } [[R_I0]], 0
; CHECK-NEXT: [[EL_ELEM01:%.*]] = extractvalue { i32, i1 } [[R_I1]], 0
; CHECK-NEXT: [[EL_ELEM02:%.*]] = extractvalue { i32, i1 } [[R_I2]], 0
; CHECK-NEXT: [[EL_UPTO0:%.*]] = insertelement <3 x i32> poison, i32 [[EL_ELEM0]], i64 0
; CHECK-NEXT: [[EL_UPTO1:%.*]] = insertelement <3 x i32> [[EL_UPTO0]], i32 [[EL_ELEM01]], i64 1
; CHECK-NEXT: [[EL:%.*]] = insertelement <3 x i32> [[EL_UPTO1]], i32 [[EL_ELEM02]], i64 2
; CHECK-NEXT: ret <3 x i32> [[EL]]
;
%r = call { <3 x i32>, <3 x i1> } @llvm.sadd.with.overflow.v3i32(<3 x i32> %b, <3 x i32> %b)
%el = extractvalue { <3 x i32>, <3 x i1> } %r, 0
ret <3 x i32> %el
}
17 changes: 17 additions & 0 deletions llvm/test/Transforms/Scalarizer/sincos.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

; Test to make sure that struct return intrinsics that are not `isTriviallyScalarizable` do not get scalarized.

define <4 x float> @test_(<4 x float> %Val) {
; CHECK-LABEL: define <4 x float> @test_(
; CHECK-SAME: <4 x float> [[VAL:%.*]]) {
; CHECK-NEXT: [[R:%.*]] = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float> [[VAL]])
; CHECK-NEXT: [[EL:%.*]] = extractvalue { <4 x float>, <4 x float> } [[R]], 0
; CHECK-NEXT: ret <4 x float> [[EL]]
;
%r = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float> %Val)
%el = extractvalue { <4 x float>, <4 x float> } %r, 0
ret <4 x float> %el
}

24 changes: 24 additions & 0 deletions llvm/test/Transforms/Scalarizer/smul_with_overflow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

define <3 x i32> @test_(<3 x i32> %a, <3 x i32> %b) {
; CHECK-LABEL: define <3 x i32> @test_(
; CHECK-SAME: <3 x i32> [[A:%.*]], <3 x i32> [[B:%.*]]) {
; CHECK-NEXT: [[B_I0:%.*]] = extractelement <3 x i32> [[B]], i64 0
; CHECK-NEXT: [[R_I0:%.*]] = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 [[B_I0]], i32 [[B_I0]])
; CHECK-NEXT: [[B_I1:%.*]] = extractelement <3 x i32> [[B]], i64 1
; CHECK-NEXT: [[R_I1:%.*]] = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 [[B_I1]], i32 [[B_I1]])
; CHECK-NEXT: [[B_I2:%.*]] = extractelement <3 x i32> [[B]], i64 2
; CHECK-NEXT: [[R_I2:%.*]] = call { i32, i1 } @llvm.smul.with.overflow.i32(i32 [[B_I2]], i32 [[B_I2]])
; CHECK-NEXT: [[EL_ELEM0:%.*]] = extractvalue { i32, i1 } [[R_I0]], 0
; CHECK-NEXT: [[EL_ELEM01:%.*]] = extractvalue { i32, i1 } [[R_I1]], 0
; CHECK-NEXT: [[EL_ELEM02:%.*]] = extractvalue { i32, i1 } [[R_I2]], 0
; CHECK-NEXT: [[EL_UPTO0:%.*]] = insertelement <3 x i32> poison, i32 [[EL_ELEM0]], i64 0
; CHECK-NEXT: [[EL_UPTO1:%.*]] = insertelement <3 x i32> [[EL_UPTO0]], i32 [[EL_ELEM01]], i64 1
; CHECK-NEXT: [[EL:%.*]] = insertelement <3 x i32> [[EL_UPTO1]], i32 [[EL_ELEM02]], i64 2
; CHECK-NEXT: ret <3 x i32> [[EL]]
;
%r = call { <3 x i32>, <3 x i1> } @llvm.smul.with.overflow.v3i32(<3 x i32> %b, <3 x i32> %b)
%el = extractvalue { <3 x i32>, <3 x i1> } %r, 0
ret <3 x i32> %el
}
24 changes: 24 additions & 0 deletions llvm/test/Transforms/Scalarizer/ssub_with_overflow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

define <3 x i32> @test_(<3 x i32> %a, <3 x i32> %b) {
; CHECK-LABEL: define <3 x i32> @test_(
; CHECK-SAME: <3 x i32> [[A:%.*]], <3 x i32> [[B:%.*]]) {
; CHECK-NEXT: [[B_I0:%.*]] = extractelement <3 x i32> [[B]], i64 0
; CHECK-NEXT: [[R_I0:%.*]] = call { i32, i1 } @llvm.ssub.with.overflow.i32(i32 [[B_I0]], i32 [[B_I0]])
; CHECK-NEXT: [[B_I1:%.*]] = extractelement <3 x i32> [[B]], i64 1
; CHECK-NEXT: [[R_I1:%.*]] = call { i32, i1 } @llvm.ssub.with.overflow.i32(i32 [[B_I1]], i32 [[B_I1]])
; CHECK-NEXT: [[B_I2:%.*]] = extractelement <3 x i32> [[B]], i64 2
; CHECK-NEXT: [[R_I2:%.*]] = call { i32, i1 } @llvm.ssub.with.overflow.i32(i32 [[B_I2]], i32 [[B_I2]])
; CHECK-NEXT: [[EL_ELEM0:%.*]] = extractvalue { i32, i1 } [[R_I0]], 0
; CHECK-NEXT: [[EL_ELEM01:%.*]] = extractvalue { i32, i1 } [[R_I1]], 0
; CHECK-NEXT: [[EL_ELEM02:%.*]] = extractvalue { i32, i1 } [[R_I2]], 0
; CHECK-NEXT: [[EL_UPTO0:%.*]] = insertelement <3 x i32> poison, i32 [[EL_ELEM0]], i64 0
; CHECK-NEXT: [[EL_UPTO1:%.*]] = insertelement <3 x i32> [[EL_UPTO0]], i32 [[EL_ELEM01]], i64 1
; CHECK-NEXT: [[EL:%.*]] = insertelement <3 x i32> [[EL_UPTO1]], i32 [[EL_ELEM02]], i64 2
; CHECK-NEXT: ret <3 x i32> [[EL]]
;
%r = call { <3 x i32>, <3 x i1> } @llvm.ssub.with.overflow.v3i32(<3 x i32> %b, <3 x i32> %b)
%el = extractvalue { <3 x i32>, <3 x i1> } %r, 0
ret <3 x i32> %el
}
16 changes: 0 additions & 16 deletions llvm/test/Transforms/Scalarizer/uadd_overflow.ll
View file
Open in desktop

This file was deleted.

24 changes: 24 additions & 0 deletions llvm/test/Transforms/Scalarizer/uadd_with_overflow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

define <3 x i32> @test_(<3 x i32> %a, <3 x i32> %b) {
; CHECK-LABEL: define <3 x i32> @test_(
; CHECK-SAME: <3 x i32> [[A:%.*]], <3 x i32> [[B:%.*]]) {
; CHECK-NEXT: [[B_I0:%.*]] = extractelement <3 x i32> [[B]], i64 0
; CHECK-NEXT: [[R_I0:%.*]] = call { i32, i1 } @llvm.uadd.with.overflow.i32(i32 [[B_I0]], i32 [[B_I0]])
; CHECK-NEXT: [[B_I1:%.*]] = extractelement <3 x i32> [[B]], i64 1
; CHECK-NEXT: [[R_I1:%.*]] = call { i32, i1 } @llvm.uadd.with.overflow.i32(i32 [[B_I1]], i32 [[B_I1]])
; CHECK-NEXT: [[B_I2:%.*]] = extractelement <3 x i32> [[B]], i64 2
; CHECK-NEXT: [[R_I2:%.*]] = call { i32, i1 } @llvm.uadd.with.overflow.i32(i32 [[B_I2]], i32 [[B_I2]])
; CHECK-NEXT: [[EL_ELEM0:%.*]] = extractvalue { i32, i1 } [[R_I0]], 0
; CHECK-NEXT: [[EL_ELEM01:%.*]] = extractvalue { i32, i1 } [[R_I1]], 0
; CHECK-NEXT: [[EL_ELEM02:%.*]] = extractvalue { i32, i1 } [[R_I2]], 0
; CHECK-NEXT: [[EL_UPTO0:%.*]] = insertelement <3 x i32> poison, i32 [[EL_ELEM0]], i64 0
; CHECK-NEXT: [[EL_UPTO1:%.*]] = insertelement <3 x i32> [[EL_UPTO0]], i32 [[EL_ELEM01]], i64 1
; CHECK-NEXT: [[EL:%.*]] = insertelement <3 x i32> [[EL_UPTO1]], i32 [[EL_ELEM02]], i64 2
; CHECK-NEXT: ret <3 x i32> [[EL]]
;
%r = call { <3 x i32>, <3 x i1> } @llvm.uadd.with.overflow.v3i32(<3 x i32> %b, <3 x i32> %b)
%el = extractvalue { <3 x i32>, <3 x i1> } %r, 0
ret <3 x i32> %el
}
24 changes: 24 additions & 0 deletions llvm/test/Transforms/Scalarizer/umul_with_overflow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

define <3 x i32> @test_(<3 x i32> %a, <3 x i32> %b) {
; CHECK-LABEL: define <3 x i32> @test_(
; CHECK-SAME: <3 x i32> [[A:%.*]], <3 x i32> [[B:%.*]]) {
; CHECK-NEXT: [[B_I0:%.*]] = extractelement <3 x i32> [[B]], i64 0
; CHECK-NEXT: [[R_I0:%.*]] = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 [[B_I0]], i32 [[B_I0]])
; CHECK-NEXT: [[B_I1:%.*]] = extractelement <3 x i32> [[B]], i64 1
; CHECK-NEXT: [[R_I1:%.*]] = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 [[B_I1]], i32 [[B_I1]])
; CHECK-NEXT: [[B_I2:%.*]] = extractelement <3 x i32> [[B]], i64 2
; CHECK-NEXT: [[R_I2:%.*]] = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 [[B_I2]], i32 [[B_I2]])
; CHECK-NEXT: [[EL_ELEM0:%.*]] = extractvalue { i32, i1 } [[R_I0]], 0
; CHECK-NEXT: [[EL_ELEM01:%.*]] = extractvalue { i32, i1 } [[R_I1]], 0
; CHECK-NEXT: [[EL_ELEM02:%.*]] = extractvalue { i32, i1 } [[R_I2]], 0
; CHECK-NEXT: [[EL_UPTO0:%.*]] = insertelement <3 x i32> poison, i32 [[EL_ELEM0]], i64 0
; CHECK-NEXT: [[EL_UPTO1:%.*]] = insertelement <3 x i32> [[EL_UPTO0]], i32 [[EL_ELEM01]], i64 1
; CHECK-NEXT: [[EL:%.*]] = insertelement <3 x i32> [[EL_UPTO1]], i32 [[EL_ELEM02]], i64 2
; CHECK-NEXT: ret <3 x i32> [[EL]]
;
%r = call { <3 x i32>, <3 x i1> } @llvm.umul.with.overflow.v3i32(<3 x i32> %b, <3 x i32> %b)
%el = extractvalue { <3 x i32>, <3 x i1> } %r, 0
ret <3 x i32> %el
}
24 changes: 24 additions & 0 deletions llvm/test/Transforms/Scalarizer/usub_with_overflow.ll
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,24 @@
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt %s -passes="function(scalarizer)" -S | FileCheck %s

define <3 x i32> @test_(<3 x i32> %a, <3 x i32> %b) {
; CHECK-LABEL: define <3 x i32> @test_(
; CHECK-SAME: <3 x i32> [[A:%.*]], <3 x i32> [[B:%.*]]) {
; CHECK-NEXT: [[B_I0:%.*]] = extractelement <3 x i32> [[B]], i64 0
; CHECK-NEXT: [[R_I0:%.*]] = call { i32, i1 } @llvm.usub.with.overflow.i32(i32 [[B_I0]], i32 [[B_I0]])
; CHECK-NEXT: [[B_I1:%.*]] = extractelement <3 x i32> [[B]], i64 1
; CHECK-NEXT: [[R_I1:%.*]] = call { i32, i1 } @llvm.usub.with.overflow.i32(i32 [[B_I1]], i32 [[B_I1]])
; CHECK-NEXT: [[B_I2:%.*]] = extractelement <3 x i32> [[B]], i64 2
; CHECK-NEXT: [[R_I2:%.*]] = call { i32, i1 } @llvm.usub.with.overflow.i32(i32 [[B_I2]], i32 [[B_I2]])
; CHECK-NEXT: [[EL_ELEM0:%.*]] = extractvalue { i32, i1 } [[R_I0]], 0
; CHECK-NEXT: [[EL_ELEM01:%.*]] = extractvalue { i32, i1 } [[R_I1]], 0
; CHECK-NEXT: [[EL_ELEM02:%.*]] = extractvalue { i32, i1 } [[R_I2]], 0
; CHECK-NEXT: [[EL_UPTO0:%.*]] = insertelement <3 x i32> poison, i32 [[EL_ELEM0]], i64 0
; CHECK-NEXT: [[EL_UPTO1:%.*]] = insertelement <3 x i32> [[EL_UPTO0]], i32 [[EL_ELEM01]], i64 1
; CHECK-NEXT: [[EL:%.*]] = insertelement <3 x i32> [[EL_UPTO1]], i32 [[EL_ELEM02]], i64 2
; CHECK-NEXT: ret <3 x i32> [[EL]]
;
%r = call { <3 x i32>, <3 x i1> } @llvm.usub.with.overflow.v3i32(<3 x i32> %b, <3 x i32> %b)
%el = extractvalue { <3 x i32>, <3 x i1> } %r, 0
ret <3 x i32> %el
}