diff options
| author | Alexey Bataev <a.bataev@hotmail.com> | 2019-08-26 17:32:45 +0000 |
|---|---|---|
| committer | Alexey Bataev <a.bataev@hotmail.com> | 2019-08-26 17:32:45 +0000 |
| commit | da8b5cc9f138b20c75f1b0592d0b046b8d5b08b6 (patch) | |
| tree | cfea28298c286930c7dc050bc94639a0fe792bb1 | |
| parent | 98f9e94e57d62170ee5c7dcd4e4d90ddb84baf2d (diff) | |
| download | bcm5719-llvm-da8b5cc9f138b20c75f1b0592d0b046b8d5b08b6.tar.gz bcm5719-llvm-da8b5cc9f138b20c75f1b0592d0b046b8d5b08b6.zip | |
[OPENMP][NVPTX]Add __kmpc_syncwarp(int32_t) function.
Summary:
Added function void __kmpc_syncwarp(int32_t) to expose it to the
compiler. It is required to fix the problem with the critical regions in
Cuda9.0+. We cannot use barrier in the critical region, but still need
to reconverge the threads in the warp after. This function allows to do
this.
Reviewers: grokos, jdoerfert
Subscribers: guansong, openmp-commits, kkwli0, caomhin
Tags: #openmp
Differential Revision: https://reviews.llvm.org/D66672
llvm-svn: 369933
| -rw-r--r-- | openmp/libomptarget/deviceRTLs/nvptx/src/interface.h | 2 | ||||
| -rw-r--r-- | openmp/libomptarget/deviceRTLs/nvptx/src/sync.cu | 10 | ||||
| -rw-r--r-- | openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h | 2 |
3 files changed, 14 insertions, 0 deletions
diff --git a/openmp/libomptarget/deviceRTLs/nvptx/src/interface.h b/openmp/libomptarget/deviceRTLs/nvptx/src/interface.h index b2a13a41f6d..ab57715592e 100644 --- a/openmp/libomptarget/deviceRTLs/nvptx/src/interface.h +++ b/openmp/libomptarget/deviceRTLs/nvptx/src/interface.h @@ -423,6 +423,8 @@ EXTERN void __kmpc_flush(kmp_Ident *loc); // vote EXTERN int32_t __kmpc_warp_active_thread_mask(); +// syncwarp +EXTERN void __kmpc_syncwarp(int32_t); // tasks EXTERN kmp_TaskDescr *__kmpc_omp_task_alloc(kmp_Ident *loc, diff --git a/openmp/libomptarget/deviceRTLs/nvptx/src/sync.cu b/openmp/libomptarget/deviceRTLs/nvptx/src/sync.cu index 4607d6a7d8a..fcfe272bc0a 100644 --- a/openmp/libomptarget/deviceRTLs/nvptx/src/sync.cu +++ b/openmp/libomptarget/deviceRTLs/nvptx/src/sync.cu @@ -11,6 +11,7 @@ //===----------------------------------------------------------------------===// #include "omptarget-nvptx.h" +#include "target_impl.h" //////////////////////////////////////////////////////////////////////////////// // KMP Ordered calls @@ -144,3 +145,12 @@ EXTERN int32_t __kmpc_warp_active_thread_mask() { PRINT0(LD_IO, "call __kmpc_warp_active_thread_mask\n"); return __ACTIVEMASK(); } + +//////////////////////////////////////////////////////////////////////////////// +// Syncwarp +//////////////////////////////////////////////////////////////////////////////// + +EXTERN void __kmpc_syncwarp(int32_t Mask) { + PRINT0(LD_IO, "call __kmpc_syncwarp\n"); + __kmpc_impl_syncwarp(Mask); +} diff --git a/openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h b/openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h index 610469527cf..a1b4c20653c 100644 --- a/openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h +++ b/openmp/libomptarget/deviceRTLs/nvptx/src/target_impl.h @@ -38,4 +38,6 @@ INLINE int __kmpc_impl_ffs(uint32_t x) { return __ffs(x); } INLINE int __kmpc_impl_popc(uint32_t x) { return __popc(x); } +INLINE void __kmpc_impl_syncwarp(int32_t Mask) { __SYNCWARP(Mask); } + #endif |
