[Openmp-commits] [PATCH] D65013: [OPENMP][NVPTX]Fix parallel level counter in Cuda 9.0.

[Alexey Bataev via Phabricator via Openmp-commits]

ABataev updated this revision to Diff 213675.
ABataev added a comment.

Rebase.


Repository:
  rOMP OpenMP

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D65013/new/

https://reviews.llvm.org/D65013

Files:
  libomptarget/deviceRTLs/nvptx/src/omptarget-nvptx.h
  libomptarget/deviceRTLs/nvptx/src/supporti.h


Index: libomptarget/deviceRTLs/nvptx/src/supporti.h
===================================================================
--- libomptarget/deviceRTLs/nvptx/src/supporti.h
+++ libomptarget/deviceRTLs/nvptx/src/supporti.h
@@ -202,25 +202,31 @@
 // Parallel level
 
 INLINE void IncParallelLevel(bool ActiveParallel) {
-  unsigned tnum = __ACTIVEMASK();
-  int leader = __ffs(tnum) - 1;
-  __SHFL_SYNC(tnum, leader, leader);
-  if (GetLaneId() == leader) {
+  unsigned Active = __ACTIVEMASK();
+  __SYNCWARP(Active);
+  unsigned LaneMaskLt;
+  asm("mov.u32 %0, %%lanemask_lt;" : "=r"(LaneMaskLt));
+  unsigned Rank = __popc(Active & LaneMaskLt);
+  if (Rank == 0) {
     parallelLevel[GetWarpId()] +=
         (1 + (ActiveParallel ? OMP_ACTIVE_PARALLEL_LEVEL : 0));
+    __threadfence();
   }
-  __SHFL_SYNC(tnum, leader, leader);
+  __SYNCWARP(Active);
 }
 
 INLINE void DecParallelLevel(bool ActiveParallel) {
-  unsigned tnum = __ACTIVEMASK();
-  int leader = __ffs(tnum) - 1;
-  __SHFL_SYNC(tnum, leader, leader);
-  if (GetLaneId() == leader) {
+  unsigned Active = __ACTIVEMASK();
+  __SYNCWARP(Active);
+  unsigned LaneMaskLt;
+  asm("mov.u32 %0, %%lanemask_lt;" : "=r"(LaneMaskLt));
+  unsigned Rank = __popc(Active & LaneMaskLt);
+  if (Rank == 0) {
     parallelLevel[GetWarpId()] -=
         (1 + (ActiveParallel ? OMP_ACTIVE_PARALLEL_LEVEL : 0));
+    __threadfence();
   }
-  __SHFL_SYNC(tnum, leader, leader);
+  __SYNCWARP(Active);
 }
 
 ////////////////////////////////////////////////////////////////////////////////
Index: libomptarget/deviceRTLs/nvptx/src/omptarget-nvptx.h
===================================================================
--- libomptarget/deviceRTLs/nvptx/src/omptarget-nvptx.h
+++ libomptarget/deviceRTLs/nvptx/src/omptarget-nvptx.h
@@ -55,11 +55,14 @@
 #define __SHFL_DOWN_SYNC(mask, var, delta, width)                              \
   __shfl_down_sync((mask), (var), (delta), (width))
 #define __ACTIVEMASK() __activemask()
+#define __SYNCWARP(Mask) __syncwarp(Mask)
 #else
 #define __SHFL_SYNC(mask, var, srcLane) __shfl((var), (srcLane))
 #define __SHFL_DOWN_SYNC(mask, var, delta, width)                              \
   __shfl_down((var), (delta), (width))
 #define __ACTIVEMASK() __ballot(1)
+// In Cuda < 9.0 no need to sync threads in warps.
+#define __SYNCWARP(Mask)
 #endif // CUDA_VERSION
 
 #define __SYNCTHREADS_N(n) asm volatile("bar.sync %0;" : : "r"(n) : "memory");


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D65013.213675.patch
Type: text/x-patch
Size: 2420 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/openmp-commits/attachments/20190806/dc3523d4/attachment.bin>