replica
/
hanchenye-llvm-project
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

OpenMP: Add convergent to more runtime functions 

Several of these other functions are probably also convergent, but
these two seem obviously convergent.
This commit is contained in:
Matt Arsenault 2019-10-27 15:09:49 -07:00 committed by Matt Arsenault
parent 40ab8ae9fb
commit 56a503bdba
2 changed files with 12 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
clang/lib/CodeGen/CGOpenMPRuntimeNVPTX.cpp
View File

@ -1817,14 +1817,14 @@ CGOpenMPRuntimeNVPTX::createNVPTXRuntimeFunction(unsigned Function) {
// Build int32_t __kmpc_warp_active_thread_mask(void);
auto *FnTy =
llvm::FunctionType::get(CGM.Int32Ty, llvm::None, /*isVarArg=*/false);
RTLFn = CGM.CreateRuntimeFunction(FnTy, "__kmpc_warp_active_thread_mask");
RTLFn = CGM.CreateConvergentRuntimeFunction(FnTy, "__kmpc_warp_active_thread_mask");
break;
}
case OMPRTL_NVPTX__kmpc_syncwarp: {
// Build void __kmpc_syncwarp(kmp_int32 Mask);
auto *FnTy =
llvm::FunctionType::get(CGM.VoidTy, CGM.Int32Ty, /*isVarArg=*/false);
RTLFn = CGM.CreateRuntimeFunction(FnTy, "__kmpc_syncwarp");
RTLFn = CGM.CreateConvergentRuntimeFunction(FnTy, "__kmpc_syncwarp");
break;
}
}

16
clang/test/OpenMP/nvptx_parallel_codegen.cpp
View File

@ -88,7 +88,7 @@ int bar(int n){
// CHECK: br label {{%?}}[[AWAIT_WORK:.+]]
//
// CHECK: [[AWAIT_WORK]]
// CHECK: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* null, i32 0) #[[#BARRIER_ATTRS:]]
// CHECK: call void @__kmpc_barrier_simple_spmd(%struct.ident_t* null, i32 0) #[[#CONVERGENT:]]
// CHECK: [[KPR:%.+]] = call i1 @__kmpc_kernel_parallel(i8** [[OMP_WORK_FN]]
// CHECK: [[KPRB:%.+]] = zext i1 [[KPR]] to i8
// store i8 [[KPRB]], i8* [[OMP_EXEC_STATUS]], align 1
@ -318,10 +318,10 @@ int bar(int n){
// CHECK: define internal void [[PARALLEL_FN4]](
// CHECK: [[A:%.+]] = alloca i[[SZ:32|64]],
// CHECK: store i[[SZ]] 45, i[[SZ]]* %a,
// CHECK: call void @__kmpc_barrier(%struct.ident_t* @{{.+}}, i32 %{{.+}}) #[[#BARRIER_ATTRS]]
// CHECK: call void @__kmpc_barrier(%struct.ident_t* @{{.+}}, i32 %{{.+}}) #[[#CONVERGENT:]]
// CHECK: ret void
// CHECK: declare void @__kmpc_barrier(%struct.ident_t*, i32) #[[#BARRIER_ATTRS]]
// CHECK: declare void @__kmpc_barrier(%struct.ident_t*, i32) #[[#CONVERGENT]]
// CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+template.+l55}}_worker()
// CHECK-LABEL: define {{.*}}void {{@__omp_offloading_.+template.+l55}}(
@ -343,7 +343,7 @@ int bar(int n){
// CHECK-LABEL: define internal void @{{.+}}(i32* noalias %{{.+}}, i32* noalias %{{.+}}, i32* dereferenceable{{.*}})
// CHECK: [[CC:%.+]] = alloca i32,
// CHECK: [[MASK:%.+]] = call i32 @__kmpc_warp_active_thread_mask()
// CHECK: [[MASK:%.+]] = call i32 @__kmpc_warp_active_thread_mask(){{$}}
// CHECK: [[TID:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.tid.x()
// CHECK: [[NUM_THREADS:%.+]] = call i32 @llvm.nvvm.read.ptx.sreg.ntid.x()
// CHECK: store i32 0, i32* [[CC]],
@ -363,11 +363,15 @@ int bar(int n){
// CHECK: store i32
// CHECK: call void @__kmpc_end_critical(
// CHECK: call void @__kmpc_syncwarp(i32 [[MASK]])
// CHECK: call void @__kmpc_syncwarp(i32 [[MASK]]){{$}}
// CHECK: [[NEW_CC_VAL:%.+]] = add nsw i32 [[CC_VAL]], 1
// CHECK: store i32 [[NEW_CC_VAL]], i32* [[CC]],
// CHECK: br label
// CHECK: attributes #[[#BARRIER_ATTRS]] = {{.*}} convergent {{.*}}
// CHECK: declare i32 @__kmpc_warp_active_thread_mask() #[[#CONVERGENT:]]
// CHECK: declare void @__kmpc_syncwarp(i32) #[[#CONVERGENT:]]
// CHECK: attributes #[[#CONVERGENT]] = {{.*}} convergent {{.*}}
#endif