AMBER Archive (2004)

Subject: Re: AMBER: pmemd and mpich - myrinet

From: Lubos Vrbka (shnek_at_tiscali.cz)
Date: Sat Apr 24 2004 - 16:18:11 CDT

bob,

> Hard to say what it is, but I would suspect some system incompatibility
> rather than it being pentium options. It is easy enough to test. Build
> pmemd with the shipped Machine.mpich_gm_ifc (with the environment variables
> fixed up) and see what that does. Look carefully at your makefile outputs.
> I recommend always building the single processor version of pmemd first too,
> since you will pick up a lot of problems with the compile process without
> the added noise of the mpi stuff (I don't think that will help a lot here,
> but I would always go that path on a machine I had any doubts about).
i've many possible combinations of pmemd binaries optimized for p3 and
xeon machines - single, shmem, mpich-p4, mpich-gm. all is built with ifc
against the same set of libraries.
i tested all p3 and xeon binaries on the xeon machines - i am loosing a
bit with p3 binaries on xeon, but it works (with one exception).
p3-mpich-gm binary works on xeon machine, but xeon-mpich-gm doesn't (you
saw the error message). that's why i was asking whether it could be
caused by some options in the makefile. from what you told me i think i
can exclude library problems, since the library is all the times the
same. if it was bad, p3-mpich-gm wouldn't work too...

i'll check at our sysadmins about the myrinet s/w. i'll also try to run
the tests on exactly the same nodes to see whether hardware is ok...

thanks for hints, 

-- 
Lubos
_@_"
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu