From: Ulf Markwardt (ulf.markwardt_at_tu-dresden.de)
Date: Thu Feb 08 2007 - 07:01:40 CST
Hi Joshua,
> If you are trying to recompile NAMD with higher levels of optimization, is
> it necessary to recompile charm with the same optimizations?
Probably not. Charm covers the communication part, this is probably not
subject to optimizations. The same holds for tcl.
On the other hand, before running time consuming NAMD simulations, I strongly
recommend you to test different versions of fftw with a couple of compiler
flags.
We have actully seen a -O0 generating a faster code than -O3 - it depends on
architecture and comiler. - And on your type of job.
BTW: Is there a NAMD-patch to use ACML or MKL or SCSL instead of fftw?
Kind regards
Ulf
-- ______________________________________________ Dr. Ulf Markwardt TU Dresden Center for Information Services and High Performance Computing (ZIH) D-01062 Dresden Germany Phone: (+49) 351/463-33640 Fax: (+49) 351/463-37773 e-mail: Ulf.Markwardt_at_tu-dresden.de WWW: http://www.tu-dresden.de/zih
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:44:22 CST