This is an archive of the discontinued LLVM Phabricator instance.

Differential D106384

[OpenMP][deviceRTLs] Update return type of function __kmpc_parallel_level
ClosedPublic

Authored by tianshilei1992 on Jul 20 2021, 11:00 AM.

Details

Reviewers
jdoerfert
Commits
rG55c65884a41c: [OpenMP][deviceRTLs] Update return type of function __kmpc_parallel_level
Summary

In deviceRTLs, the parallel level is stored in a shared variable of type uint8_t.
__kmpc_parallel_level currently returns a 16-bit interger. This patch first
changes the return type of the function to uint8_t, same as the shared variable,
and then corrects function type which was updated in D105955.

Diff Detail

Event Timeline

tianshilei1992 created this revision.Jul 20 2021, 11:00 AM
Herald added subscribers: guansong, yaxunl. · View Herald TranscriptJul 20 2021, 11:00 AM
tianshilei1992 requested review of this revision.Jul 20 2021, 11:00 AM
Herald added a reviewer: jdoerfert. · View Herald TranscriptJul 20 2021, 11:00 AM
Herald added projects: Restricted Project, Restricted Project. · View Herald Transcript
Herald added subscribers: llvm-commits, openmp-commits, sstefan1. · View Herald Transcript
tianshilei1992 updated this revision to Diff 360211.Jul 20 2021, 11:01 AM

update

Harbormaster completed remote builds in B115151: Diff 360211.Jul 20 2021, 12:33 PM
jdoerfert accepted this revision.Jul 20 2021, 12:42 PM

LG

This revision is now accepted and ready to land.Jul 20 2021, 12:42 PM
This revision was landed with ongoing or failed builds.Jul 20 2021, 12:45 PM
Closed by commit rG55c65884a41c: [OpenMP][deviceRTLs] Update return type of function __kmpc_parallel_level (authored by tianshilei1992). · Explain Why
This revision was automatically updated to reflect the committed changes.
tianshilei1992 added a commit: rG55c65884a41c: [OpenMP][deviceRTLs] Update return type of function __kmpc_parallel_level.

Revision Contents

PathSize
llvm/
include/
llvm/
Frontend/
OpenMP/
2 lines
openmp/
libomptarget/
deviceRTLs/
common/
src/
2 lines

Diff 360210

llvm/include/llvm/Frontend/OpenMP/OMPKinds.def

Show First 20 Lines • Show All 427 Lines • ▼ Show 20 Lines
__OMP_RTL(__kmpc_shuffle_int64, false, Int64, Int64, Int16, Int16) __OMP_RTL(__kmpc_shuffle_int64, false, Int64, Int64, Int16, Int16)
__OMP_RTL(__kmpc_alloc_shared, false, VoidPtr, SizeTy) __OMP_RTL(__kmpc_alloc_shared, false, VoidPtr, SizeTy)
__OMP_RTL(__kmpc_free_shared, false, Void, VoidPtr) __OMP_RTL(__kmpc_free_shared, false, Void, VoidPtr)
__OMP_RTL(__kmpc_begin_sharing_variables, false, Void, VoidPtrPtrPtr, SizeTy) __OMP_RTL(__kmpc_begin_sharing_variables, false, Void, VoidPtrPtrPtr, SizeTy)
__OMP_RTL(__kmpc_end_sharing_variables, false, Void, ) __OMP_RTL(__kmpc_end_sharing_variables, false, Void, )
__OMP_RTL(__kmpc_get_shared_variables, false, Void, VoidPtrPtrPtr) __OMP_RTL(__kmpc_get_shared_variables, false, Void, VoidPtrPtrPtr)
__OMP_RTL(__kmpc_parallel_level, false, Int16, IdentPtr, Int32) __OMP_RTL(__kmpc_parallel_level, false, Int8, )
__OMP_RTL(__kmpc_is_spmd_exec_mode, false, Int8, ) __OMP_RTL(__kmpc_is_spmd_exec_mode, false, Int8, )
__OMP_RTL(__kmpc_barrier_simple_spmd, false, Void, IdentPtr, Int32) __OMP_RTL(__kmpc_barrier_simple_spmd, false, Void, IdentPtr, Int32)
__OMP_RTL(__kmpc_warp_active_thread_mask, false, LanemaskTy,) __OMP_RTL(__kmpc_warp_active_thread_mask, false, LanemaskTy,)
__OMP_RTL(__kmpc_syncwarp, false, Void, LanemaskTy) __OMP_RTL(__kmpc_syncwarp, false, Void, LanemaskTy)
__OMP_RTL(__last, false, Void, ) __OMP_RTL(__last, false, Void, )
▲ Show 20 Lines • Show All 721 Lines • Show Last 20 Lines

openmp/libomptarget/deviceRTLs/common/src/parallel.cu

Show First 20 Lines • Show All 232 Lines • ▼ Show 20 LinesEXTERN void __kmpc_end_serialized_parallel(kmp_Ident *loc,
omptarget_nvptx_threadPrivateContext->SetTopLevelTaskDescr( omptarget_nvptx_threadPrivateContext->SetTopLevelTaskDescr(
threadId, currTaskDescr->GetPrevTaskDescr()); threadId, currTaskDescr->GetPrevTaskDescr());
// free // free
SafeFree(currTaskDescr, "new seq parallel task"); SafeFree(currTaskDescr, "new seq parallel task");
currTaskDescr = getMyTopTaskDescriptor(threadId); currTaskDescr = getMyTopTaskDescriptor(threadId);
currTaskDescr->RestoreLoopData(); currTaskDescr->RestoreLoopData();
} }
EXTERN uint16_t __kmpc_parallel_level() { EXTERN uint8_t __kmpc_parallel_level() {
return parallelLevel[GetWarpId()] & (OMP_ACTIVE_PARALLEL_LEVEL - 1); return parallelLevel[GetWarpId()] & (OMP_ACTIVE_PARALLEL_LEVEL - 1);
} }
// This kmpc call returns the thread id across all teams. It's value is // This kmpc call returns the thread id across all teams. It's value is
// cached by the compiler and used when calling the runtime. On nvptx // cached by the compiler and used when calling the runtime. On nvptx
// it's cheap to recalculate this value so we never use the result // it's cheap to recalculate this value so we never use the result
// of this call. // of this call.
EXTERN int32_t __kmpc_global_thread_num(kmp_Ident *loc) { EXTERN int32_t __kmpc_global_thread_num(kmp_Ident *loc) {
▲ Show 20 Lines • Show All 106 Lines • Show Last 20 Lines