maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[OPENMP][NVPTX]Use __kmpc_data_sharing_coalesced_push_stack function. 

Coalesced memory access requires use of the new function
`__kmpc_data_sharing_coalesced_push_stack` instead of the
`__kmpc_data_sharing_push_stack`.

llvm-svn: 345991
This commit is contained in:
Alexey Bataev 2018-11-02 16:08:31 +00:00
parent 391b19c78a
commit 1fc1f8e819
3 changed files with 20 additions and 19 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
View File

@ -84,9 +84,9 @@ enum OpenMPRTLFunctionNVPTX {
OMPRTL_NVPTX__kmpc_data_sharing_init_stack,
/// Call to void __kmpc_data_sharing_init_stack_spmd();
OMPRTL_NVPTX__kmpc_data_sharing_init_stack_spmd,
/// Call to void* __kmpc_data_sharing_push_stack(size_t size,
/// Call to void* __kmpc_data_sharing_coalesced_push_stack(size_t size,
/// int16_t UseSharedMemory);
OMPRTL_NVPTX__kmpc_data_sharing_push_stack,
OMPRTL_NVPTX__kmpc_data_sharing_coalesced_push_stack,
/// Call to void __kmpc_data_sharing_pop_stack(void *a);
OMPRTL_NVPTX__kmpc_data_sharing_pop_stack,
/// Call to void __kmpc_begin_sharing_variables(void ***args,
@ -1745,14 +1745,14 @@ CGOpenMPRuntimeNVPTX::createNVPTXRuntimeFunction(unsigned Function) {
CGM.CreateRuntimeFunction(FnTy, "__kmpc_data_sharing_init_stack_spmd");
break;
}
case OMPRTL_NVPTX__kmpc_data_sharing_push_stack: {
// Build void *__kmpc_data_sharing_push_stack(size_t size,
case OMPRTL_NVPTX__kmpc_data_sharing_coalesced_push_stack: {
// Build void *__kmpc_data_sharing_coalesced_push_stack(size_t size,
// int16_t UseSharedMemory);
llvm::Type *TypeParams[] = {CGM.SizeTy, CGM.Int16Ty};
auto *FnTy =
llvm::FunctionType::get(CGM.VoidPtrTy, TypeParams, /*isVarArg=*/false);
RTLFn = CGM.CreateRuntimeFunction(
FnTy, /*Name=*/"__kmpc_data_sharing_push_stack");
FnTy, /*Name=*/"__kmpc_data_sharing_coalesced_push_stack");
break;
}
case OMPRTL_NVPTX__kmpc_data_sharing_pop_stack: {
@ -2105,10 +2105,10 @@ void CGOpenMPRuntimeNVPTX::emitGenericVarsProlog(CodeGenFunction &CGF,
// the user, for now, default to global.
llvm::Value *GlobalRecordSizeArg[] = {
Size, CGF.Builder.getInt16(/*UseSharedMemory=*/0)};
llvm::Value *GlobalRecValue =
CGF.EmitRuntimeCall(createNVPTXRuntimeFunction(
OMPRTL_NVPTX__kmpc_data_sharing_push_stack),
GlobalRecordSizeArg);
llvm::Value *GlobalRecValue = CGF.EmitRuntimeCall(
createNVPTXRuntimeFunction(
OMPRTL_NVPTX__kmpc_data_sharing_coalesced_push_stack),
GlobalRecordSizeArg);
GlobalRecCastAddr = Bld.CreatePointerBitCastOrAddrSpaceCast(
GlobalRecValue, GlobalRecPtrTy);
CGF.EmitBlock(ExitBB);
@ -2182,10 +2182,10 @@ void CGOpenMPRuntimeNVPTX::emitGenericVarsProlog(CodeGenFunction &CGF,
llvm::Value *GlobalRecordSizeArg[] = {
llvm::ConstantInt::get(CGM.SizeTy, GlobalRecordSize),
CGF.Builder.getInt16(/*UseSharedMemory=*/0)};
llvm::Value *GlobalRecValue =
CGF.EmitRuntimeCall(createNVPTXRuntimeFunction(
OMPRTL_NVPTX__kmpc_data_sharing_push_stack),
GlobalRecordSizeArg);
llvm::Value *GlobalRecValue = CGF.EmitRuntimeCall(
createNVPTXRuntimeFunction(
OMPRTL_NVPTX__kmpc_data_sharing_coalesced_push_stack),
GlobalRecordSizeArg);
GlobalRecCastAddr = Bld.CreatePointerBitCastOrAddrSpaceCast(
GlobalRecValue, GlobalRecPtrTy);
I->getSecond().GlobalRecordAddr = GlobalRecValue;
@ -2282,7 +2282,8 @@ void CGOpenMPRuntimeNVPTX::emitGenericVarsProlog(CodeGenFunction &CGF,
llvm::Value *GlobalRecordSizeArg[] = {
Size, CGF.Builder.getInt16(/*UseSharedMemory=*/0)};
llvm::Value *GlobalRecValue = CGF.EmitRuntimeCall(
createNVPTXRuntimeFunction(OMPRTL_NVPTX__kmpc_data_sharing_push_stack),
createNVPTXRuntimeFunction(
OMPRTL_NVPTX__kmpc_data_sharing_coalesced_push_stack),
GlobalRecordSizeArg);
llvm::Value *GlobalRecCastAddr = Bld.CreatePointerBitCastOrAddrSpaceCast(
GlobalRecValue, CGF.ConvertTypeForMem(VD->getType())->getPointerTo());

8
clang/test/OpenMP/declare_target_codegen_globalization.cpp
View File

@ -10,7 +10,7 @@ int bar() {
}
// CHECK: define weak void @__omp_offloading_{{.*}}maini1{{.*}}_l[[@LINE+5]](i32* dereferenceable{{.*}})
// CHECK-NOT: @__kmpc_data_sharing_push_stack
// CHECK-NOT: @__kmpc_data_sharing_coalesced_push_stack
int maini1() {
int a;
@ -24,7 +24,7 @@ int maini1() {
// parallel region
// CHECK: define {{.*}}void @{{.*}}(i32* noalias {{.*}}, i32* noalias {{.*}}, i32* dereferenceable{{.*}})
// CHECK-NOT: call i8* @__kmpc_data_sharing_push_stack(
// CHECK-NOT: call i8* @__kmpc_data_sharing_coalesced_push_stack(
// CHECK: [[B_ADDR:%.+]] = alloca i32,
// CHECK: call {{.*}}[[FOO:@.*foo.*]](i32* dereferenceable{{.*}} [[B_ADDR]])
// CHECK: call {{.*}}[[BAR:@.*bar.*]]()
@ -32,7 +32,7 @@ int maini1() {
// CHECK: ret void
// CHECK: define {{.*}}[[FOO]](i32* dereferenceable{{.*}})
// CHECK-NOT: @__kmpc_data_sharing_push_stack
// CHECK-NOT: @__kmpc_data_sharing_coalesced_push_stack
// CHECK: define {{.*}}[[BAR]]()
// CHECK: alloca i32,
@ -41,7 +41,7 @@ int maini1() {
// CHECK: [[IS_SPMD:%.+]] = icmp ne i8 [[RES]], 0
// CHECK: br i1 [[IS_SPMD]], label
// CHECK: br label
// CHECK: [[RES:%.+]] = call i8* @__kmpc_data_sharing_push_stack(i64 128, i16 0)
// CHECK: [[RES:%.+]] = call i8* @__kmpc_data_sharing_coalesced_push_stack(i64 128, i16 0)
// CHECK: [[GLOBALS:%.+]] = bitcast i8* [[RES]] to [[GLOBAL_ST:%.+]]*
// CHECK: br label
// CHECK: [[ITEMS:%.+]] = phi [[GLOBAL_ST]]* [ null, {{.+}} ], [ [[GLOBALS]], {{.+}} ]

2
clang/test/OpenMP/nvptx_target_codegen.cpp
View File

@ -566,7 +566,7 @@ int baz(int f, double &a) {
// CHECK: br i1 [[IS_SPMD]], label
// CHECK: br label
// CHECK: [[SIZE:%.+]] = select i1 [[IS_TTD]], i{{64|32}} 4, i{{64|32}} 128
// CHECK: [[PTR:%.+]] = call i8* @__kmpc_data_sharing_push_stack(i{{64|32}} [[SIZE]], i16 0)
// CHECK: [[PTR:%.+]] = call i8* @__kmpc_data_sharing_coalesced_push_stack(i{{64|32}} [[SIZE]], i16 0)
// CHECK: [[REC_ADDR:%.+]] = bitcast i8* [[PTR]] to [[GLOBAL_ST:%.+]]*
// CHECK: br label
// CHECK: [[ITEMS:%.+]] = phi [[GLOBAL_ST]]* [ null, {{.+}} ], [ [[REC_ADDR]], {{.+}} ]