Re: OpenMP parallelization in AMBER6

From: Carlos Simmerling <>
Date: Mon 13 Mar 2000 15:04:44 -0500

OpenMP may be better, but MPI does not take twice as long as reported in my
experience.. Can you be specific:

what Irix version
what machine
what processor
how many threads you were trying to use
was the machine empty, or fully loaded?

These all influence the parallel timings. I've done quite a bit of performance
testing of AMBER5 vs AMBER6 for parallel sander and can share my
experiences if you give me the details.

I also find that the MPI single processor code runs in nearly the same
time as the non-parallel compiled version. Have you tested that?

My understanding ifrom SGI is that they had done an OpenMP port, but on a less
recent version of sander. They are now working on AMBER6, or so I have
been told.

Marios Philippopoulos wrote:

> Hi everyone,
> I have been told that MPI parallelization is not optimal on SGI
> machines, whereas OpenMP is. From my own recent experience, I have found
> that my jobs take about twice as long to complete as suggested by the
> actual cpu time reported in the output file. I just checked the AMBER6
> manual, and in the INSTALL.parallel file (in the "Shared Memory"
> section) it is stated that:
> "Under development is an OpenMP version which should be portable to a
> variety
> of shared memory multiprocessor machines."
> Would anyone know when the OpenMP implementation might be ready for use?
> Also, any comments on any of the above are welcome.
> Thanks,
> Marios
> --
> Marios Philippopoulos, PhD
> Structural Biology and Biochemistry Programme
> Hospital for Sick Children
> 555 University Avenue
> Toronto Ontario Canada M5G 1X8

Carlos L. Simmerling, Ph.D.
Assistant Professor           Phone: (631) 632-1336  NOTE NEW AREA CODE
Department of Chemistry       Fax:   (631) 632-7960
SUNY at Stony Brook           Web:
Stony Brook, NY 11794-3400    E-mail:
Received on Mon Mar 13 2000 - 12:04:44 PST
Custom Search