Hi Ryan,
It varies but is about 8% performance regression on average. If you have 980s you have to use cuda 6.5 so it's probably simplest on a shared resource to just install cuda 6.5 and not worry about the performance regression.
All the best
Ross
> On Feb 5, 2015, at 12:48 PM, Ryan Novosielski <novosirj.ca.rutgers.edu> wrote:
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi all,
>
> I've seen online and on this mailing list to use CUDA 5.0 instead of
> 6.5 for performance reasons. What I've not seen is anything more
> specific about what the performance penalty might be. I think that it
> is supposed to be minor (though I have a preference for running with
> whatever is fastest if it's that little work to do it).
>
> Thanks!
>
> - --
> ____ *Note: UMDNJ is now Rutgers-Biomedical and Health Sciences*
> || \\UTGERS |---------------------*O*---------------------
> ||_// Biomedical | Ryan Novosielski - Senior Technologist
> || \\ and Health | novosirj.rutgers.edu - 973/972.0922 (2x0922)
> || \\ Sciences | OIRT/High Perf & Res Comp - MSB C630, Newark
> `'
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1
>
> iEYEARECAAYFAlTT1wMACgkQmb+gadEcsb7aOgCfV45x43uA3EVuZkZvBUt9q4Ut
> XLwAoIcMA+wdtr8bBwBWqlJ5tvHHOdj9
> =iRf6
> -----END PGP SIGNATURE-----
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Feb 05 2015 - 13:00:04 PST
