On Wed, Apr 18, 2018, Aravind Ravichandran wrote:
> I am performing TMD and have amber version 14. When I use a single
> node/32 processors for performing TMD simulation, simulations speed up
> to 3ns/day, whereas any increase in the nodes/processors slows down the
> simulation to 0.2ns/day.
This is not unusual. Sander rarely scales beyond a single node, or a
few dozen MPI threads. You could try a sample run without targetting,
but I suspect that any job that goes off-node or beyond 32 threads is
likely to show decreased, rather than increased, performance.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 18 2018 - 06:30:03 PDT
