[Openmp-commits] [openmp] cf37a94 - [openmp] Add amdgpu impl missed from D112153

[Jon Chesterfield via Openmp-commits]

Author: Jon Chesterfield
Date: 2021-10-28T00:55:53+01:00
New Revision: cf37a94c1e42ce5b6eff272098c678c021c46c3b

URL: https://github.com/llvm/llvm-project/commit/cf37a94c1e42ce5b6eff272098c678c021c46c3b
DIFF: https://github.com/llvm/llvm-project/commit/cf37a94c1e42ce5b6eff272098c678c021c46c3b.diff

LOG: [openmp] Add amdgpu impl missed from D112153

Added: 
    

Modified: 
    openmp/libomptarget/DeviceRTL/src/Synchronization.cpp

Removed: 
    


################################################################################
diff  --git a/openmp/libomptarget/DeviceRTL/src/Synchronization.cpp b/openmp/libomptarget/DeviceRTL/src/Synchronization.cpp
index 33e2194b25f3..931dffcaa131 100644
--- a/openmp/libomptarget/DeviceRTL/src/Synchronization.cpp
+++ b/openmp/libomptarget/DeviceRTL/src/Synchronization.cpp
@@ -192,6 +192,7 @@ void syncWarp(__kmpc_impl_lanemask_t) {
 }
 
 void syncThreads() { __builtin_amdgcn_s_barrier(); }
+void syncThreadsAligned() { syncThreads(); }
 
 // TODO: Don't have wavefront lane locks. Possibly can't have them.
 void unsetLock(omp_lock_t *) { __builtin_trap(); }