https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105201
Bug ID: 105201
Summary: cuLaunchKernel error with subsequent empty target
regions
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: libgomp
Assignee: unassigned at gcc dot gnu.org
Reporter: marcel at codesourcery dot com
CC: jakub at gcc dot gnu.org
Target Milestone: ---
Hi all,
I observed a
"libgomp: cuLaunchKernel error: invalid resource handle"
for the following testcase (compiled with x86_64-linux and run with nvptx
offloading (different machines with different kinds of GPUs)):
#include <omp.h>
int
main (int argc, char *const *argv)
{
__builtin_printf ("omp_get_num_devices (): %d\n", omp_get_num_devices ());
for (int i=0; i < 2; i++)
#pragma omp target device (i)
;
#pragma omp target device (0)
;
return 0;
}
The error occurred when using two devices at least (for (n)one device the test
passes).
Marcel