> That is true for the main (Mac OS X) partition. Things are slightly
> different on the Linux partition. Since the compiler locations and
> versions are somewhat different, and the IBM compilers gave me fits,
> I decided to go for options that I thought would be easiest to use
> right off - the lam that was packaged with Amber10, and gfortran,
> which I was more used to working with.
The biggest problem there is that using any other MPI but the packaged
MVAPICH deal that supports the InfiniBand hardware is probably not going
to give you access to the IB interconnect, and then scaling to multiple
nodes will be really, really poor. gfortran might not be your biggest
problem.
Persist! Get the real compilers and, in particular, the interconnect,
to work for you, and you will see excellent performance and very good
scaling. You can get AMBER and PMEMD to work without, as a sort of
"proof of concept", but if you're dreaming of actually doing real work
with anything more than two nodes...
--
Atro Tossavainen (Mr.) / The Institute of Biotechnology at
Systems Analyst, Techno-Amish & / the University of Helsinki, Finland,
+358-9-19158939 UNIX Dinosaur / employs me, but my opinions are my own.
< URL : http : / / www . helsinki . fi / %7E atossava / > NO FILE ATTACHMENTS
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
to majordomo.scripps.edu
Received on Sun Jun 01 2008 - 06:08:16 PDT
