| Message ID | 20190109001004.GA29783@delia |
|---|---|
| State | New |
| Headers | show |
| Series | [committed,nvptx,libgomp] Don't launch with num_workers == 0 | expand |
diff --git a/libgomp/plugin/plugin-nvptx.c b/libgomp/plugin/plugin-nvptx.c index 572d9ef8d5c..60553bdf3bd 100644 --- a/libgomp/plugin/plugin-nvptx.c +++ b/libgomp/plugin/plugin-nvptx.c @@ -1272,6 +1272,7 @@ nvptx_exec (void (*fn), size_t mapnum, void **hostaddrs, void **devaddrs, ? vectors : dims[GOMP_DIM_VECTOR]); workers = blocks / actual_vectors; + workers = MAX (workers, 1); } for (i = 0; i != GOMP_DIM_MAX; i++)
Hi, When using a compiler build with: ... +#define PTX_DEFAULT_VECTOR_LENGTH PTX_CTA_SIZE +#define PTX_MAX_VECTOR_LENGTH PTX_CTA_SIZE ... and running the libgomp testsuite, we run into an execution failure in parallel-loop-1.c, due to a cuda launch failure: ... nvptx_exec: kernel f6_none_none$_omp_fn$0: launch gangs=480, workers=0, \ vectors=1024 libgomp: cuLaunchKernel error: invalid argument ... because workers == 0. The workers variable is set to 0 here in nvptx_exec: ... workers = blocks / actual_vectors; ... because actual_vectors is 1024, and blocks is 768: ... cuOccupancyMaxPotentialBlockSize: grid = 10, block = 768 ... Fix this by ensuring that workers is at least one. Committed to trunk. Thanks, - Tom [nvptx, libgomp] Don't launch with num_workers == 0 2019-01-08 Tom de Vries <tdevries@suse.de> * plugin/plugin-nvptx.c (nvptx_exec): Make sure to launch with at least one worker. --- libgomp/plugin/plugin-nvptx.c | 1 + 1 file changed, 1 insertion(+)