AMBER Archive (2004)

Subject: Re: AMBER: running on Itanium machines

From: David A. Case (case_at_scripps.edu)
Date: Fri Jan 09 2004 - 10:43:48 CST


On Thu, Jan 08, 2004, northrop_at_chem.ucla.edu wrote:
>
> Has anyone had trouble running simulations on Itanium machines using Intel
> compilers? I have a number of simulations that don't run correctly on our
> Intel IA64 cluster. They run fine on our local IA32 cluster, but when run
> on the IA64 I've noticed the following problems: Jobs run in implicit
> solvent don't output any energy information. Instead, in the output file,
> the energy and temperature terms are all just question marks. Jobs run in
> explicit solvent result in wildly fluctuating potential energies. The
> potential energy terms (when using constant volume or constant pressure)
> fluctuate by thousands of Kcal. The same jobs run on our local cluster
> fluctuate by less than 10 kcal.

We need more information about what OS, which version of the compilers you are
using, and what compiler options you used. I would suggest compiling with
lower optimization (definitely not -O3), if you have not already tried that.
Various versions of efc (aka ifort) have failed to compile amber correctly
with -O3 set.

(Amber does work on some[many] IA64 machines using Intel compilers; which is
why we need more info.)

..hope this helps...dac

-- 

================================================================== David A. Case | e-mail: case_at_scripps.edu Dept. of Molecular Biology, TPC15 | fax: +1-858-784-8896 The Scripps Research Institute | phone: +1-858-784-9768 10550 N. Torrey Pines Rd. | home page: La Jolla CA 92037 USA | http://www.scripps.edu/case ================================================================== ----------------------------------------------------------------------- The AMBER Mail Reflector To post, send mail to amber_at_scripps.edu To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu