Re: [AMBER] test failed but still work

From: Jason Swails <jason.swails.gmail.com>
Date: Wed, 15 Jul 2015 10:32:44 -0700

I really should read these in order :).

On Wed, Jul 15, 2015 at 11:35 AM, David A Case <david.case.rutgers.edu>
wrote:

> On Wed, Jul 15, 2015, Albert wrote:
> > Hello:
> >
> > I've compiled AMber with command:
> >
> > ./configure -intelmpi intel
>
> You need *both* the -mpi and the -intelmpi flags.
>
> (Jason: can we fix this? have -intelmpi imply -mpi?)
>

​I think this is already the case... This is the snippet from configure2
that handles these flags:

while [ $# -gt 0 ]; do
    case "$1" in
        -mpi) mpi='yes'; mpinab='mpi';mtkpp='' ;;
        -intelmpi) mpi='yes'; intelmpi='yes'; mpinab='mpi';mtkpp='';
rism='no'; export MPICC="mpiicc" ;;

So -intelmpi already sets mpi, mpinab, ... etc. So I don't think both -mpi
and -intelmpi is necessary (but I don't have Intel MPI, so I've never been
able to test it).

All the best,
Jason

-
​-​
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jul 15 2015 - 11:00:02 PDT
Custom Search