Re: NAMD 2.9 CUDA multicore, single machine

From: Aron Broom (broomsday_at_gmail.com)
Date: Fri Jun 22 2012 - 13:49:30 CDT

You want to download the multicore binary, and then the command is quite
simple:

namd2 +p3 +idlepoll simulation.conf

On Fri, Jun 22, 2012 at 2:04 PM, Brett T. Hannigan
<brett_at_thehannigans.org>wrote:

> When using NAMD 2.8 CUDA, I could run on my local machine using multiple
> cores by using charmrun.
> *charmrun ++local +p3 namd2 +idlepoll simulation.conf*
> However, when I downloaded NAMD 2.9 and tried the same thing, it dies
> saying
> "FATAL ERROR: Unknown command-line option ++local"
>
> It looks like the charmrun included in the NAMD 2.9 CUDA is a simple
> script as described for the MPI version, while the charmrun in the 2.8 CUDA
> version I downloaded is an actual binary. Any suggestions on what the
> proper thing to do here is?
>
> For the time being, I downloaded the non-CUDA NAMD 2.9, which has a binary
> version of charmrun, and I'm trying using that charmrun with the CUDA
> namd2, but I'm not sure if it's proper to mix and match like that...
>
> --
> ******************************************
> Brett T. Hannigan
> 8011 Ardleigh St.
> Philadelphia, PA 19118
> brett.hannigan_at_gmail.com
> ******************************************
>

-- 
Aron Broom M.Sc
PhD Student
Department of Chemistry
University of Waterloo

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:42 CST