Skip to content

Commit f2be466

Browse files
dhruvachakCopilot
andauthored
Update offload/plugins-nextgen/amdgpu/src/rtl.cpp
Co-authored-by: Copilot <[email protected]>
1 parent d6d480f commit f2be466

File tree

1 file changed

+1
-1
lines changed
  • offload/plugins-nextgen/amdgpu/src

1 file changed

+1
-1
lines changed

offload/plugins-nextgen/amdgpu/src/rtl.cpp

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1294,7 +1294,7 @@ struct AMDGPUKernelTy : public GenericKernelTy {
12941294

12951295
uint64_t PreferredNumBlocks = TripCountNumBlocks;
12961296
// Occupancy-based setting overrides block reuse.
1297-
if (OMPX_GenericSPMDOccupancyBasedOpt) {
1297+
if (OMPX_GenericSPMDOccupancyBasedOpt && NumTeamsEnvVar == 0 && NumTeamsClause[0] == 0) {
12981298
PreferredNumBlocks =
12991299
std::min(PreferredNumBlocks,
13001300
OptimizeNumTeamsBaseOccupancy(GenericDevice, NumThreads));

0 commit comments

Comments
 (0)