[AMBER] Problem (perhaps) with new precision model

From: Fabrício Bracht <bracht.iq.ufrj.br>
Date: Tue, 11 Sep 2012 20:23:17 -0300

Hello. I am running MD calculations under SPFP using the latest
patches on Amber12 with pmemd.cuda. I have divided my 100 ns
simulation into 10 parts and some of the simulations end with an
error:

cudaMemcpy GpuBuffer::Download failed unspecified launch failure

I am using a GTX580 to do this. This sort of problem never happened
before. My system is comprised of a zinc bound protein. The parameters
I obtained from the MTK++ routine. Should I go back to using SPDP?
Thank you
Fabrício Bracht

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Sep 11 2012 - 16:30:02 PDT
Custom Search