I tend to do that too although I should say this is with Gromacs which,
by default, probes the hardware where it is built on. With AMBER I
just always compiled without looking into all this SSE et al. stuff.
Another option would be to try and optimise for the target architecture
(executables with different names or separate directories) and then run
through a wrapper. I once had a simple shell script looking
into /proc/cpuinfo to guess what executable needed to be run. There are
probably more sophisticated ways of doing this.
Cheers,
Hannes.
On Wed, 14 Jan 2015 13:26:33 +0100
Jan-Philip Gehrcke <jgehrcke.googlemail.com> wrote:
> Let me propose another quite simple and yet efficient solution on a
> heterogeneous cluster:
>
> Compile on the machine with the least hardware features w/o changing
> any of the configure options (i.e. using xHost).
>
> The resulting build works and uses the intersection of advanced
> hardware features on all machines in your cluster (SSE 4.1 in your
> case). This is what I have been doing many times now, and it really
> saves a lot of work.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jan 14 2015 - 05:30:10 PST
