diff options
| author | Alexey Bataev <a.bataev@hotmail.com> | 2018-04-06 16:03:36 +0000 |
|---|---|---|
| committer | Alexey Bataev <a.bataev@hotmail.com> | 2018-04-06 16:03:36 +0000 |
| commit | e290ec02c7b38929c6d78d5c026a1db312641add (patch) | |
| tree | 77d75026c4c3bc59a488274be2da33199a3d662c /clang/test/OpenMP/nvptx_target_codegen.cpp | |
| parent | 63ae5579e7cabf519db0e5c8cd69d4683b34b6e8 (diff) | |
| download | bcm5719-llvm-e290ec02c7b38929c6d78d5c026a1db312641add.tar.gz bcm5719-llvm-e290ec02c7b38929c6d78d5c026a1db312641add.zip | |
[OPENMP, NVPTX] Fix codegen for the teams reduction.
Added NUW flags for all the add|mul|sub operations + replaced sdiv by udiv
as we operate on unsigned values only (addresses, converted to integers)
llvm-svn: 329411
Diffstat (limited to 'clang/test/OpenMP/nvptx_target_codegen.cpp')
| -rw-r--r-- | clang/test/OpenMP/nvptx_target_codegen.cpp | 24 |
1 files changed, 12 insertions, 12 deletions
diff --git a/clang/test/OpenMP/nvptx_target_codegen.cpp b/clang/test/OpenMP/nvptx_target_codegen.cpp index 23b40e10c4c..70f3973acd4 100644 --- a/clang/test/OpenMP/nvptx_target_codegen.cpp +++ b/clang/test/OpenMP/nvptx_target_codegen.cpp @@ -68,7 +68,7 @@ int foo(int n) { // CHECK-DAG: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x() // CHECK-DAG: [[NTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[WS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK-DAG: [[TH_LIMIT:%.+]] = sub i32 [[NTH]], [[WS]] + // CHECK-DAG: [[TH_LIMIT:%.+]] = sub nuw i32 [[NTH]], [[WS]] // CHECK: [[IS_WORKER:%.+]] = icmp ult i32 [[TID]], [[TH_LIMIT]] // CHECK: br i1 [[IS_WORKER]], label {{%?}}[[WORKER:.+]], label {{%?}}[[CHECK_MASTER:.+]] // @@ -86,7 +86,7 @@ int foo(int n) { // CHECK: [[MASTER]] // CHECK-DAG: [[MNTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]] + // CHECK: [[MTMP1:%.+]] = sub nuw i32 [[MNTH]], [[MWS]] // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]] // CHECK: br label {{%?}}[[TERMINATE:.+]] // @@ -144,7 +144,7 @@ int foo(int n) { // CHECK-DAG: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x() // CHECK-DAG: [[NTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[WS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK-DAG: [[TH_LIMIT:%.+]] = sub i32 [[NTH]], [[WS]] + // CHECK-DAG: [[TH_LIMIT:%.+]] = sub nuw i32 [[NTH]], [[WS]] // CHECK: [[IS_WORKER:%.+]] = icmp ult i32 [[TID]], [[TH_LIMIT]] // CHECK: br i1 [[IS_WORKER]], label {{%?}}[[WORKER:.+]], label {{%?}}[[CHECK_MASTER:.+]] // @@ -162,7 +162,7 @@ int foo(int n) { // CHECK: [[MASTER]] // CHECK-DAG: [[MNTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]] + // CHECK: [[MTMP1:%.+]] = sub nuw i32 [[MNTH]], [[MWS]] // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]] // CHECK: load i16, i16* [[AA_CADDR]], // CHECK: br label {{%?}}[[TERMINATE:.+]] @@ -245,7 +245,7 @@ int foo(int n) { // CHECK-DAG: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x() // CHECK-DAG: [[NTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[WS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK-DAG: [[TH_LIMIT:%.+]] = sub i32 [[NTH]], [[WS]] + // CHECK-DAG: [[TH_LIMIT:%.+]] = sub nuw i32 [[NTH]], [[WS]] // CHECK: [[IS_WORKER:%.+]] = icmp ult i32 [[TID]], [[TH_LIMIT]] // CHECK: br i1 [[IS_WORKER]], label {{%?}}[[WORKER:.+]], label {{%?}}[[CHECK_MASTER:.+]] // @@ -263,7 +263,7 @@ int foo(int n) { // CHECK: [[MASTER]] // CHECK-DAG: [[MNTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]] + // CHECK: [[MTMP1:%.+]] = sub nuw i32 [[MNTH]], [[MWS]] // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]] // // Use captures. @@ -414,7 +414,7 @@ int bar(int n){ // CHECK-DAG: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x() // CHECK-DAG: [[NTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[WS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK-DAG: [[TH_LIMIT:%.+]] = sub i32 [[NTH]], [[WS]] + // CHECK-DAG: [[TH_LIMIT:%.+]] = sub nuw i32 [[NTH]], [[WS]] // CHECK: [[IS_WORKER:%.+]] = icmp ult i32 [[TID]], [[TH_LIMIT]] // CHECK: br i1 [[IS_WORKER]], label {{%?}}[[WORKER:.+]], label {{%?}}[[CHECK_MASTER:.+]] // @@ -432,7 +432,7 @@ int bar(int n){ // CHECK: [[MASTER]] // CHECK-DAG: [[MNTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]] + // CHECK: [[MTMP1:%.+]] = sub nuw i32 [[MNTH]], [[MWS]] // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]] // CHECK-64-DAG: load i32, i32* [[REF_A]] // CHECK-32-DAG: load i32, i32* [[LOCAL_A]] @@ -503,7 +503,7 @@ int bar(int n){ // CHECK-DAG: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x() // CHECK-DAG: [[NTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[WS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK-DAG: [[TH_LIMIT:%.+]] = sub i32 [[NTH]], [[WS]] + // CHECK-DAG: [[TH_LIMIT:%.+]] = sub nuw i32 [[NTH]], [[WS]] // CHECK: [[IS_WORKER:%.+]] = icmp ult i32 [[TID]], [[TH_LIMIT]] // CHECK: br i1 [[IS_WORKER]], label {{%?}}[[WORKER:.+]], label {{%?}}[[CHECK_MASTER:.+]] // @@ -521,7 +521,7 @@ int bar(int n){ // CHECK: [[MASTER]] // CHECK-DAG: [[MNTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]] + // CHECK: [[MTMP1:%.+]] = sub nuw i32 [[MNTH]], [[MWS]] // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]] // Use captures. // CHECK-DAG: getelementptr inbounds [[S1]], [[S1]]* [[REF_THIS]], i32 0, i32 0 @@ -587,7 +587,7 @@ int bar(int n){ // CHECK-DAG: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x() // CHECK-DAG: [[NTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[WS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK-DAG: [[TH_LIMIT:%.+]] = sub i32 [[NTH]], [[WS]] + // CHECK-DAG: [[TH_LIMIT:%.+]] = sub nuw i32 [[NTH]], [[WS]] // CHECK: [[IS_WORKER:%.+]] = icmp ult i32 [[TID]], [[TH_LIMIT]] // CHECK: br i1 [[IS_WORKER]], label {{%?}}[[WORKER:.+]], label {{%?}}[[CHECK_MASTER:.+]] // @@ -605,7 +605,7 @@ int bar(int n){ // CHECK: [[MASTER]] // CHECK-DAG: [[MNTH:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x() // CHECK-DAG: [[MWS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.warpsize() - // CHECK: [[MTMP1:%.+]] = sub i32 [[MNTH]], [[MWS]] + // CHECK: [[MTMP1:%.+]] = sub nuw i32 [[MNTH]], [[MWS]] // CHECK: call void @__kmpc_kernel_init(i32 [[MTMP1]] // // CHECK-64-DAG: load i32, i32* [[REF_A]] |
