Re: [AMBER-Developers] questions about compiler flags

From: case <case.biomaps.rutgers.edu>
Date: Tue, 2 Feb 2010 08:21:54 -0500

On Tue, Feb 02, 2010, Mengjuei Hsieh wrote:
>
> If performance is really a concern for certain programs, how about
> we release some well-optimized executables for them on some major
> platforms?

No thanks...I don't want to go down that road. (1) We don't have a
mechanism for releasing binaries of for updating them after bugfixes.
(2) I don't know how to have a "universal" MPI code linked in, and
having something that doesn't match the hardware is likely to impact
performance far more than compiler options. (3) We would need to purchase
"commercial" compiler licenses. (4) If you look at the number of options
in Bob's config_data directory, you can see that some very large number of
executables would be required.

On another subject: Can you folks look at the pg_pgb test case? I get the
same result on two different platforms, but both disagree with the save
output.

...thx...dac


_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Tue Feb 02 2010 - 05:30:03 PST
Custom Search