From: David Hardy (dhardy_at_ks.uiuc.edu)
Date: Wed Sep 10 2014 - 14:46:22 CDT
To increase the number of patches defined by NAMD, you can use some combination of the twoAwayX, twoAwayY, or twoAwayZ config file options. These subdivide the default patch size along the respective directions. Enabling all of these options together will increase the patch count by a factor of 8. See the user guide for more details.
On Sep 9, 2014, at 6:59 PM, Siva Dasetty <sdasett_at_g.clemson.edu> wrote:
> Dear All,
> I am running a simulation of a small protein (4K atoms) in implicit solvent using cuda enabled NAMD and I only get 2-3 ns/day computational speed when I am using a single node with 10 (or less) processors and 2 gpus (k20's).
> However if I try to increase the number of nodes, I get the following error "CUDA-enabled NAMD requires at least one patch per process."
> and if I increase the number of processors >10 in a single node, I also get the following note "MPI_ABORT was invoked on rank 4 in communicator MPI_COMM_WORLD
> with errorcode 1." along with the above fatal error.
> I tried to follow previous threads in the archive list but couldn't understand anything much about this (http://www.ks.uiuc.edu/Research/namd/mailing_list/namd-l/11376.html).
> Is this because my system size is too small ? or am I doing something absurd while executing the job?
> I have seen better benchmarks reported by NVDIA (http://www.nvidia.com/docs/IO/122634/NAMD-benchmark-report.pdf) and others (http://www.ks.uiuc.edu/Research/gpu/files/ICMS-Edinburgh/GB.pdf).
-- David J. Hardy, Ph.D. Theoretical and Computational Biophysics Beckman Institute, University of Illinois dhardy_at_ks.uiuc.edu http://www.ks.uiuc.edu/~dhardy/
This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:50 CST