Neil
unread,Apr 7, 2021, 3:50:15 PM4/7/21You do not have permission to delete messages in this group
Sign in to report message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to
Dear c.l.f.,
I'm interested in adapting an existing Fortran program to make use of
GPUs by adding directives. I've played around a little with adding
OpenACC directives and, using nvfortran, have succeeded in obtaining a
small speedup to one part of the program (and some large slowdowns to
other bits). Before I go further, I was wondering if anyone could
advise me about the pros and cons of using OpenACC vs. OpenMP 4+ for
this task. In particular, the most important issue seems to be what
direction Fortran compilers are going to take over the next few years.
Thanks very much for any help,
Neil.