[Openmp-commits] [openmp] r348772 - [OPENMP][NVPTX]Revert __kmpc_shuffle_int64 to its original form.

Alexey Bataev via Openmp-commits openmp-commits at lists.llvm.org
Mon Dec 10 08:50:37 PST 2018


Author: abataev
Date: Mon Dec 10 08:50:36 2018
New Revision: 348772

URL: http://llvm.org/viewvc/llvm-project?rev=348772&view=rev
Log:
[OPENMP][NVPTX]Revert __kmpc_shuffle_int64 to its original form.

Summary:
Use the original shuffle implementation for __kmpc_shuffle_int64 since
default implementation uses the same implementation.

Reviewers: gtbercea

Subscribers: guansong, caomhin, openmp-commits

Differential Revision: https://reviews.llvm.org/D55514

Modified:
    openmp/trunk/libomptarget/deviceRTLs/nvptx/src/reduction.cu

Modified: openmp/trunk/libomptarget/deviceRTLs/nvptx/src/reduction.cu
URL: http://llvm.org/viewvc/llvm-project/openmp/trunk/libomptarget/deviceRTLs/nvptx/src/reduction.cu?rev=348772&r1=348771&r2=348772&view=diff
==============================================================================
--- openmp/trunk/libomptarget/deviceRTLs/nvptx/src/reduction.cu (original)
+++ openmp/trunk/libomptarget/deviceRTLs/nvptx/src/reduction.cu Mon Dec 10 08:50:36 2018
@@ -76,17 +76,12 @@ EXTERN int32_t __kmpc_shuffle_int32(int3
 }
 
 EXTERN int64_t __kmpc_shuffle_int64(int64_t val, int16_t delta, int16_t size) {
-#if defined(CUDART_VERSION) && CUDART_VERSION >= 9000
-  return __SHFL_DOWN_SYNC(0xFFFFFFFFFFFFFFFFLL, (long long)val, (unsigned)delta,
-                          (int)size);
-#else
    int lo, hi;
    asm volatile("mov.b64 {%0,%1}, %2;" : "=r"(lo), "=r"(hi) : "l"(val));
    hi = __SHFL_DOWN_SYNC(0xFFFFFFFF, hi, delta, size);
    lo = __SHFL_DOWN_SYNC(0xFFFFFFFF, lo, delta, size);
    asm volatile("mov.b64 %0, {%1,%2};" : "=l"(val) : "r"(lo), "r"(hi));
    return val;
-#endif
 }
 
 static INLINE void gpu_regular_warp_reduce(void *reduce_data,




More information about the Openmp-commits mailing list