[clang] 946b326 - [InstSimplify] sdiv a (1 srem b) --> a
Nikita Popov via cfe-commits
cfe-commits at lists.llvm.org
Mon Apr 24 05:37:17 PDT 2023
Author: Siyuan Zhu
Date: 2023-04-24T14:37:07+02:00
New Revision: 946b32680311f43a349d0199f9e286f385cd9847
URL: https://github.com/llvm/llvm-project/commit/946b32680311f43a349d0199f9e286f385cd9847
DIFF: https://github.com/llvm/llvm-project/commit/946b32680311f43a349d0199f9e286f385cd9847.diff
LOG: [InstSimplify] sdiv a (1 srem b) --> a
Extend the existing fold for division by zero or one to use known
bits, so it catches additional patterns like division by
(1 srem b).
Fixes https://github.com/llvm/llvm-project/issues/62163.
Differential Revision: https://reviews.llvm.org/D149001
Added:
Modified:
clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
clang/test/OpenMP/for_reduction_task_codegen.cpp
clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
clang/test/OpenMP/parallel_reduction_task_codegen.cpp
clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
clang/test/OpenMP/sections_reduction_task_codegen.cpp
clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
llvm/lib/Analysis/InstructionSimplify.cpp
llvm/test/Transforms/InstCombine/zext-or-icmp.ll
llvm/test/Transforms/InstSimplify/div.ll
Removed:
################################################################################
diff --git a/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
index 7cd29b2f70574..de9b18bbb5fba 100644
--- a/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/distribute_parallel_for_reduction_task_codegen.cpp
@@ -568,22 +568,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/for_reduction_task_codegen.cpp b/clang/test/OpenMP/for_reduction_task_codegen.cpp
index 687aa973dde27..48587ce9a34b6 100644
--- a/clang/test/OpenMP/for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/for_reduction_task_codegen.cpp
@@ -481,22 +481,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
index 7596fee02ffd6..2770409411af2 100644
--- a/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_for_reduction_task_codegen.cpp
@@ -476,22 +476,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
index a4cd31d332df6..b48892e7eafca 100644
--- a/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_master_reduction_task_codegen.cpp
@@ -431,22 +431,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_reduction_task_codegen.cpp
index d6a30678b1638..0a4f21fc9bf0b 100644
--- a/clang/test/OpenMP/parallel_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_reduction_task_codegen.cpp
@@ -422,22 +422,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp b/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
index e2347ef9836d1..bc48f493eeeee 100644
--- a/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/parallel_sections_reduction_task_codegen.cpp
@@ -464,22 +464,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/sections_reduction_task_codegen.cpp b/clang/test/OpenMP/sections_reduction_task_codegen.cpp
index 29965f0e70637..2af379d7c0068 100644
--- a/clang/test/OpenMP/sections_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/sections_reduction_task_codegen.cpp
@@ -469,22 +469,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
index 141e80fcbb3d8..f6b9bfa0d026f 100644
--- a/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/target_parallel_for_reduction_task_codegen.cpp
@@ -489,22 +489,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp b/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
index b7891d8a93aa9..f1f257c5dac94 100644
--- a/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/target_parallel_reduction_task_codegen.cpp
@@ -435,22 +435,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
index 6d17c48f51707..df9cd7731f856 100644
--- a/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/target_teams_distribute_parallel_for_reduction_task_codegen.cpp
@@ -828,22 +828,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp b/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
index f8f09ca593159..846aa737453b9 100644
--- a/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
+++ b/clang/test/OpenMP/teams_distribute_parallel_for_reduction_task_codegen.cpp
@@ -832,22 +832,20 @@ int main(int argc, char **argv) {
// CHECK1-NEXT: [[TMP28:%.*]] = ptrtoint ptr [[ARRAYIDX3_I]] to i64
// CHECK1-NEXT: [[TMP29:%.*]] = ptrtoint ptr [[TMP20]] to i64
// CHECK1-NEXT: [[TMP30:%.*]] = sub i64 [[TMP28]], [[TMP29]]
-// CHECK1-NEXT: [[TMP31:%.*]] = sdiv exact i64 [[TMP30]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP32:%.*]] = add nuw i64 [[TMP31]], 1
-// CHECK1-NEXT: [[TMP33:%.*]] = mul nuw i64 [[TMP32]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: store i64 [[TMP32]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
-// CHECK1-NEXT: [[TMP34:%.*]] = load ptr, ptr [[TMP12]], align 8
-// CHECK1-NEXT: [[TMP35:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP34]], ptr [[TMP20]])
-// CHECK1-NEXT: [[TMP36:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP31:%.*]] = add nuw i64 [[TMP30]], 1
+// CHECK1-NEXT: [[TMP32:%.*]] = mul nuw i64 [[TMP31]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
+// CHECK1-NEXT: store i64 [[TMP31]], ptr @{{reduction_size[.].+[.]}}, align 8, !noalias !12
+// CHECK1-NEXT: [[TMP33:%.*]] = load ptr, ptr [[TMP12]], align 8
+// CHECK1-NEXT: [[TMP34:%.*]] = call ptr @__kmpc_task_reduction_get_th_data(i32 [[TMP16]], ptr [[TMP33]], ptr [[TMP20]])
+// CHECK1-NEXT: [[TMP35:%.*]] = getelementptr inbounds [[STRUCT_ANON]], ptr [[TMP9]], i32 0, i32 2
+// CHECK1-NEXT: [[TMP36:%.*]] = load ptr, ptr [[TMP35]], align 8
// CHECK1-NEXT: [[TMP37:%.*]] = load ptr, ptr [[TMP36]], align 8
-// CHECK1-NEXT: [[TMP38:%.*]] = load ptr, ptr [[TMP37]], align 8
-// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP38]] to i64
-// CHECK1-NEXT: [[TMP40:%.*]] = ptrtoint ptr [[TMP20]] to i64
-// CHECK1-NEXT: [[TMP41:%.*]] = sub i64 [[TMP39]], [[TMP40]]
-// CHECK1-NEXT: [[TMP42:%.*]] = sdiv exact i64 [[TMP41]], ptrtoint (ptr getelementptr (i8, ptr null, i32 1) to i64)
-// CHECK1-NEXT: [[TMP43:%.*]] = getelementptr i8, ptr [[TMP35]], i64 [[TMP42]]
+// CHECK1-NEXT: [[TMP38:%.*]] = ptrtoint ptr [[TMP37]] to i64
+// CHECK1-NEXT: [[TMP39:%.*]] = ptrtoint ptr [[TMP20]] to i64
+// CHECK1-NEXT: [[TMP40:%.*]] = sub i64 [[TMP38]], [[TMP39]]
+// CHECK1-NEXT: [[TMP41:%.*]] = getelementptr i8, ptr [[TMP34]], i64 [[TMP40]]
// CHECK1-NEXT: store ptr [[TMP4_I]], ptr [[TMP_I]], align 8, !noalias !12
-// CHECK1-NEXT: store ptr [[TMP43]], ptr [[TMP4_I]], align 8, !noalias !12
+// CHECK1-NEXT: store ptr [[TMP41]], ptr [[TMP4_I]], align 8, !noalias !12
// CHECK1-NEXT: ret i32 0
//
//
diff --git a/llvm/lib/Analysis/InstructionSimplify.cpp b/llvm/lib/Analysis/InstructionSimplify.cpp
index addc724872158..0097c618eaa5e 100644
--- a/llvm/lib/Analysis/InstructionSimplify.cpp
+++ b/llvm/lib/Analysis/InstructionSimplify.cpp
@@ -1128,17 +1128,17 @@ static Value *simplifyDivRem(Instruction::BinaryOps Opcode, Value *Op0,
// X / 1 -> X
// X % 1 -> 0
- // If this is a boolean op (single-bit element type), we can't have
- // division-by-zero or remainder-by-zero, so assume the divisor is 1.
- // Similarly, if we're zero-extending a boolean divisor, then assume it's a 1.
- Value *X;
- if (match(Op1, m_One()) || Ty->isIntOrIntVectorTy(1) ||
- (match(Op1, m_ZExt(m_Value(X))) && X->getType()->isIntOrIntVectorTy(1)))
+ // If the divisor can only be zero or one, we can't have division-by-zero
+ // or remainder-by-zero, so assume the divisor is 1.
+ // e.g. 1, zext (i8 X), sdiv X (Y and 1)
+ KnownBits Known = computeKnownBits(Op1, Q.DL, 0, Q.AC, Q.CxtI, Q.DT);
+ if (Known.countMinLeadingZeros() == Known.getBitWidth() - 1)
return IsDiv ? Op0 : Constant::getNullValue(Ty);
// If X * Y does not overflow, then:
// X * Y / Y -> X
// X * Y % Y -> 0
+ Value *X;
if (match(Op0, m_c_Mul(m_Value(X), m_Specific(Op1)))) {
auto *Mul = cast<OverflowingBinaryOperator>(Op0);
// The multiplication can't overflow if it is defined not to, or if
diff --git a/llvm/test/Transforms/InstCombine/zext-or-icmp.ll b/llvm/test/Transforms/InstCombine/zext-or-icmp.ll
index 05759b2ca6db9..655c2e7c083a9 100644
--- a/llvm/test/Transforms/InstCombine/zext-or-icmp.ll
+++ b/llvm/test/Transforms/InstCombine/zext-or-icmp.ll
@@ -248,7 +248,7 @@ define i1 @PR51762(ptr %i, i32 %t0, i16 %t1, ptr %p, ptr %d, ptr %f, i32 %p2, i1
; CHECK-NEXT: store i32 [[ADD]], ptr [[F]], align 4
; CHECK-NEXT: [[REM18:%.*]] = srem i32 [[LOR_EXT]], [[ADD]]
; CHECK-NEXT: [[CONV19:%.*]] = zext i32 [[REM18]] to i64
-; CHECK-NEXT: store i32 0, ptr [[D]], align 8
+; CHECK-NEXT: store i32 [[SROA38]], ptr [[D]], align 8
; CHECK-NEXT: [[R:%.*]] = icmp ult i64 [[INSERT_INSERT41]], [[CONV19]]
; CHECK-NEXT: call void @llvm.assume(i1 [[R]])
; CHECK-NEXT: ret i1 true
diff --git a/llvm/test/Transforms/InstSimplify/div.ll b/llvm/test/Transforms/InstSimplify/div.ll
index a5f23015043d0..a379e1ec9efe2 100644
--- a/llvm/test/Transforms/InstSimplify/div.ll
+++ b/llvm/test/Transforms/InstSimplify/div.ll
@@ -437,9 +437,7 @@ define <2 x i8> @udiv_exact_trailing_zeros_nonuniform_vector(<2 x i8> %x) {
define i32 @sdiv_one_srem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @sdiv_one_srem_divisor(
-; CHECK-NEXT: [[SREM:%.*]] = srem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[SDIV:%.*]] = sdiv i32 [[A:%.*]], [[SREM]]
-; CHECK-NEXT: ret i32 [[SDIV]]
+; CHECK-NEXT: ret i32 [[A:%.*]]
;
%srem = srem i32 1, %b
%sdiv = sdiv i32 %a, %srem
@@ -448,9 +446,7 @@ define i32 @sdiv_one_srem_divisor(i32 %a, i32 %b) {
define i32 @sdiv_one_urem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @sdiv_one_urem_divisor(
-; CHECK-NEXT: [[UREM:%.*]] = urem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[SDIV:%.*]] = sdiv i32 [[A:%.*]], [[UREM]]
-; CHECK-NEXT: ret i32 [[SDIV]]
+; CHECK-NEXT: ret i32 [[A:%.*]]
;
%urem = urem i32 1, %b
%sdiv = sdiv i32 %a, %urem
@@ -459,9 +455,7 @@ define i32 @sdiv_one_urem_divisor(i32 %a, i32 %b) {
define i32 @udiv_one_srem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @udiv_one_srem_divisor(
-; CHECK-NEXT: [[SREM:%.*]] = srem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[UDIV:%.*]] = udiv i32 [[A:%.*]], [[SREM]]
-; CHECK-NEXT: ret i32 [[UDIV]]
+; CHECK-NEXT: ret i32 [[A:%.*]]
;
%srem = srem i32 1, %b
%udiv = udiv i32 %a, %srem
@@ -470,9 +464,7 @@ define i32 @udiv_one_srem_divisor(i32 %a, i32 %b) {
define i32 @udiv_one_urem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @udiv_one_urem_divisor(
-; CHECK-NEXT: [[UREM:%.*]] = urem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[UDIV:%.*]] = udiv i32 [[A:%.*]], [[UREM]]
-; CHECK-NEXT: ret i32 [[UDIV]]
+; CHECK-NEXT: ret i32 [[A:%.*]]
;
%urem = urem i32 1, %b
%udiv = udiv i32 %a, %urem
@@ -481,9 +473,7 @@ define i32 @udiv_one_urem_divisor(i32 %a, i32 %b) {
define i32 @srem_one_srem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @srem_one_srem_divisor(
-; CHECK-NEXT: [[SREM:%.*]] = srem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[SREM1:%.*]] = srem i32 [[A:%.*]], [[SREM]]
-; CHECK-NEXT: ret i32 [[SREM1]]
+; CHECK-NEXT: ret i32 0
;
%srem = srem i32 1, %b
%srem1 = srem i32 %a, %srem
@@ -492,9 +482,7 @@ define i32 @srem_one_srem_divisor(i32 %a, i32 %b) {
define i32 @urem_one_srem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @urem_one_srem_divisor(
-; CHECK-NEXT: [[SREM:%.*]] = srem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[UREM:%.*]] = urem i32 [[A:%.*]], [[SREM]]
-; CHECK-NEXT: ret i32 [[UREM]]
+; CHECK-NEXT: ret i32 0
;
%srem = srem i32 1, %b
%urem = urem i32 %a, %srem
@@ -503,9 +491,7 @@ define i32 @urem_one_srem_divisor(i32 %a, i32 %b) {
define i32 @srem_one_urem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @srem_one_urem_divisor(
-; CHECK-NEXT: [[UREM:%.*]] = urem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[SREM:%.*]] = srem i32 [[A:%.*]], [[UREM]]
-; CHECK-NEXT: ret i32 [[SREM]]
+; CHECK-NEXT: ret i32 0
;
%urem = urem i32 1, %b
%srem = srem i32 %a, %urem
@@ -514,9 +500,7 @@ define i32 @srem_one_urem_divisor(i32 %a, i32 %b) {
define i32 @urem_one_urem_divisor(i32 %a, i32 %b) {
; CHECK-LABEL: @urem_one_urem_divisor(
-; CHECK-NEXT: [[UREM:%.*]] = urem i32 1, [[B:%.*]]
-; CHECK-NEXT: [[UREM1:%.*]] = urem i32 [[A:%.*]], [[UREM]]
-; CHECK-NEXT: ret i32 [[UREM1]]
+; CHECK-NEXT: ret i32 0
;
%urem = urem i32 1, %b
%urem1 = urem i32 %a, %urem
@@ -525,9 +509,7 @@ define i32 @urem_one_urem_divisor(i32 %a, i32 %b) {
define <2 x i8> @sdiv_one_vec_srem_divisor(<2 x i8> %a, <2 x i8> %b) {
; CHECK-LABEL: @sdiv_one_vec_srem_divisor(
-; CHECK-NEXT: [[SREM:%.*]] = srem <2 x i8> <i8 1, i8 1>, [[B:%.*]]
-; CHECK-NEXT: [[SDIV:%.*]] = sdiv <2 x i8> [[A:%.*]], [[SREM]]
-; CHECK-NEXT: ret <2 x i8> [[SDIV]]
+; CHECK-NEXT: ret <2 x i8> [[A:%.*]]
;
%srem = srem <2 x i8> <i8 1, i8 1>, %b
%sdiv = sdiv <2 x i8> %a, %srem
@@ -536,9 +518,7 @@ define <2 x i8> @sdiv_one_vec_srem_divisor(<2 x i8> %a, <2 x i8> %b) {
define i32 @sdiv_and_one_divisor(i32 %x, i32 %y) {
; CHECK-LABEL: @sdiv_and_one_divisor(
-; CHECK-NEXT: [[AND:%.*]] = and i32 [[X:%.*]], 1
-; CHECK-NEXT: [[RES:%.*]] = sdiv i32 [[Y:%.*]], [[AND]]
-; CHECK-NEXT: ret i32 [[RES]]
+; CHECK-NEXT: ret i32 [[Y:%.*]]
;
%and = and i32 %x, 1
%res = sdiv i32 %y, %and
@@ -547,9 +527,7 @@ define i32 @sdiv_and_one_divisor(i32 %x, i32 %y) {
define <2 x i8> @sdiv_and_one_vec_divisor(<2 x i8> %x, <2 x i8> %y) {
; CHECK-LABEL: @sdiv_and_one_vec_divisor(
-; CHECK-NEXT: [[AND:%.*]] = and <2 x i8> [[X:%.*]], <i8 1, i8 1>
-; CHECK-NEXT: [[RES:%.*]] = sdiv <2 x i8> [[Y:%.*]], [[AND]]
-; CHECK-NEXT: ret <2 x i8> [[RES]]
+; CHECK-NEXT: ret <2 x i8> [[Y:%.*]]
;
%and = and <2 x i8> %x, <i8 1, i8 1>
%res = sdiv <2 x i8> %y, %and
@@ -558,10 +536,7 @@ define <2 x i8> @sdiv_and_one_vec_divisor(<2 x i8> %x, <2 x i8> %y) {
define i32 @sdiv_neg_or_divisor(i32 %x, i32 %y) {
; CHECK-LABEL: @sdiv_neg_or_divisor(
-; CHECK-NEXT: [[OR:%.*]] = or i32 [[X:%.*]], -2
-; CHECK-NEXT: [[NEG:%.*]] = xor i32 [[OR]], -1
-; CHECK-NEXT: [[RES:%.*]] = sdiv i32 [[Y:%.*]], [[NEG]]
-; CHECK-NEXT: ret i32 [[RES]]
+; CHECK-NEXT: ret i32 [[Y:%.*]]
;
%or = or i32 %x, -2
%neg = xor i32 %or, -1
More information about the cfe-commits
mailing list