[Scalarizer] Make `*_with_overflow` intrinsics scalarizable (#126815)
Addresses issue #126809
- Made `uadd_with_overflow`, `sadd_with_overflow`, `usub_with_overflow`,
`ssub_with_overflow`, `umul_with_overflow`, and `smul_with_overflow`
trivially scalarizable in `isTriviallyScalarizable()` from
`VectorUtils.cpp`
- Renamed and updated the test `Scalarizer/uadd_overflow.ll` to
`Scalarizer/uadd_with_overflow.ll` to check that `uadd_with_overflow`
gets scalarized
- Added a test `Scalarizer/sincos.ll` to ensure the bug fix #113625
still works
diff --git a/llvm/lib/Analysis/VectorUtils.cpp b/llvm/lib/Analysis/VectorUtils.cpp
index ad80e45..53be7fc 100644
--- a/llvm/lib/Analysis/VectorUtils.cpp
+++ b/llvm/lib/Analysis/VectorUtils.cpp
@@ -125,6 +125,12 @@
// 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;
diff --git a/llvm/test/Transforms/Scalarizer/sadd_with_overflow.ll b/llvm/test/Transforms/Scalarizer/sadd_with_overflow.ll
new file mode 100644
index 0000000..1e5c503
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/sadd_with_overflow.ll
@@ -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
+}
diff --git a/llvm/test/Transforms/Scalarizer/sincos.ll b/llvm/test/Transforms/Scalarizer/sincos.ll
new file mode 100644
index 0000000..8db4ba3
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/sincos.ll
@@ -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
+}
+
diff --git a/llvm/test/Transforms/Scalarizer/smul_with_overflow.ll b/llvm/test/Transforms/Scalarizer/smul_with_overflow.ll
new file mode 100644
index 0000000..c934077
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/smul_with_overflow.ll
@@ -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
+}
diff --git a/llvm/test/Transforms/Scalarizer/ssub_with_overflow.ll b/llvm/test/Transforms/Scalarizer/ssub_with_overflow.ll
new file mode 100644
index 0000000..3654e20
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/ssub_with_overflow.ll
@@ -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
+}
diff --git a/llvm/test/Transforms/Scalarizer/uadd_overflow.ll b/llvm/test/Transforms/Scalarizer/uadd_overflow.ll
deleted file mode 100644
index 3909445..0000000
--- a/llvm/test/Transforms/Scalarizer/uadd_overflow.ll
+++ /dev/null
@@ -1,16 +0,0 @@
-; 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 <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: [[R:%.*]] = call { <3 x i32>, <3 x i1> } @llvm.uadd.with.overflow.v3i32(<3 x i32> [[B]], <3 x i32> [[B]])
-; CHECK-NEXT: [[EL:%.*]] = extractvalue { <3 x i32>, <3 x i1> } [[R]], 0
-; 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
-}
diff --git a/llvm/test/Transforms/Scalarizer/uadd_with_overflow.ll b/llvm/test/Transforms/Scalarizer/uadd_with_overflow.ll
new file mode 100644
index 0000000..1c4b612
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/uadd_with_overflow.ll
@@ -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
+}
diff --git a/llvm/test/Transforms/Scalarizer/umul_with_overflow.ll b/llvm/test/Transforms/Scalarizer/umul_with_overflow.ll
new file mode 100644
index 0000000..ac1ca11
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/umul_with_overflow.ll
@@ -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
+}
diff --git a/llvm/test/Transforms/Scalarizer/usub_with_overflow.ll b/llvm/test/Transforms/Scalarizer/usub_with_overflow.ll
new file mode 100644
index 0000000..fe1b530
--- /dev/null
+++ b/llvm/test/Transforms/Scalarizer/usub_with_overflow.ll
@@ -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
+}
