On Mon, Mar 20, 2017 at 10:08 PM, Scott Brozell
<sbrozell.rci.rutgers.edu> wrote:
> I just pushed a fix for -mkl and cpptraj
Thanks Scott - it's been a while since I've tested MKL. (clearly).
> and i noticed that cpptraj is always using parallel mkl (via -mkl without
> any qualifications) as opposed to -mkl=sequential. Is that intentional ?
Yes. Threaded shared memory parallelism in cpptraj is at the level of
Actions and Analyses, so this should provide at least a minor speedup
for BLAS/LAPACK routines (I think). It has been a while since I've
tested though.
> Note also that the two threads of mkl support in our configure
> MKL_HOME and -mkl should probably be handled in a more unified manner -
> potential post release future work.
Agreed.
-Dan
>
> scott
>
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
--
-------------------------
Daniel R. Roe
Laboratory of Computational Biology
National Institutes of Health, NHLBI
5635 Fishers Ln, Rm T900
Rockville MD, 20852
https://www.lobos.nih.gov/lcb
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Mar 21 2017 - 08:00:02 PDT
