Re: NCL and openmpi

From: Kyle Griffin <kgriffin_at_nyahnyahspammersnyahnyah>
Date: Thu May 03 2012 - 09:51:15 MDT

Micah -

I had briefly looked into this recently, and found a previous NCL-talk post
referring to one person's project that attempted to build some OpenMPI
functionality into NCL. While I have yet to download or try any of this
code, a number of the source files are available for it. The link to the
project is:

http://code.google.com/p/mpi4ncl/

and code is available under the source tab. If you decide to try it and
have some success, I'm sure some here on this list would be interested to
hear. I doubt the performance will be comparable to the version that Mary
has mentioned, but could give you a starting point for now.

Kyle
-----------------------------------------------------
Kyle S. Griffin
Dept. of Atmospheric and Environmental Sciences
University at Albany, SUNY
1400 Washington Ave, Albany, NY 12222
Office: ES-218 Email: kgriffin@albany.edu
http://www.atmos.albany.edu/student/kgriffin/

On Thu, May 3, 2012 at 10:45 AM, Mary Haley <haley@ucar.edu> wrote:

> Hi Micah,
>
> We don't have OpenMPI in NCL, but we are working with Argonne National
> Laboratory and other DOE labs to parallelize components of NCL. This new
> version will be called "ParNCL" and we hope to release a beta version later
> this year.
>
> We are interested in hearing from users what analyses and/or file I/O
> functionality they use in NCL that they find really slow. We are currently
> prioritizing which functions and features to focus on first, and user
> feedback in this area would be very valuable.
>
> --Mary
>
> On May 2, 2012, at 9:13 AM, Micah Sklut wrote:
>
> > Hi NCL folks,
> >
> > I have been using a super old version of NCL - 4.3.1
> > I haven't updated, because I haven't had a need to.
> >
> > I am interested in running NCL in MPI mode.
> > If I remember correctly from the email list, recent versions of NCL have
> this capability?
> >
> > I am looking for documentation or direction on how to compile/run NCL
> with OpenMPI, if the capability exists?
> > I am looking for anyway to increase efficiency of processing data and
> images.
> > Any help is appreciated.
> >
> > Thank you,
> >
> > --
> > Micah Sklut
> >
> >
> > _______________________________________________
> > ncl-talk mailing list
> > List instructions, subscriber options, unsubscribe:
> > http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu May 3 09:51:47 2012

This archive was generated by hypermail 2.1.8 : Thu May 10 2012 - 16:57:50 MDT