I could not find the CUDA API to choose L1/shared memory setting in CUDA Fortran document. May I ask if it has been implemented yet in PGI 10.9.
And what is the default configuration in CUDA Fortran.
Thanks,
Tuan
I could not find the CUDA API to choose L1/shared memory setting in CUDA Fortran document. May I ask if it has been implemented yet in PGI 10.9.
And what is the default configuration in CUDA Fortran.
Thanks,
Tuan
Does any one have some information?
Thanks,
Tuan