Re: [AMBER] Test results for amber-cuda, single node, single GPU, Tesla C2070

From: Jason Swails <jason.swails.gmail.com>
Date: Wed, 25 May 2011 22:12:22 -0400

On Wed, May 25, 2011 at 7:18 PM, Ross Walker <ross.rosswalker.co.uk> wrote:

> > > Yes it does. Line 45 of $AMBERHOME/src/Makefile
> > >
> > > cuda_parallel: configured_cuda configured_parallel clean $(NETCDFLIB)
> > > .echo "Starting installation of ${AMBER} (cuda parallel) at
> > `date`".
> > > cd pmemd && $(MAKE) cuda_parallel
> > >
> > > Something smells fishy with your copy of AMBER 11 to me if it is
> > missing
> > > this.
> > >
> >
> > Could be unpatched. I don't think we had cuda_parallel at Amber11
> > release, right?
>
> But how would he get any parallel version compiled??? - Or maybe it is just
> the serial version linked with MPI. Ugh...
>

I don't think the OP ever built pmemd.cuda.MPI or ever tried outside of
getting the "no cuda_parallel rule" message. The other errors he was
getting building pmemd.cuda (I didn't read his previous email carefully
enough to realize he was trying to build serial cuda) were simply because he
had leftover files from building pmemd.MPI lying around. This gives you
exactly his error. Thus, here's what I think is happening (and it fits all
of the symptoms).

The Amber11 directory is not fully patched (up to bugfix 11 or 12,
whatever). Thus, there is no parallel target.

*Maybe* he tried building amber11 + AmberTools 1.5 without running
AT15_Amber11.py first, which will cause the netcdf.mod file *not* to get
copied, possibly giving rise to NetCDF errors. However, as no error message
was shown and I don't see how the proposed addition to the config.h can
possibly affect the success of the build process, I can't speculate with any
kind of certainty.

The OP did make serial, make clean, make parallel, make cuda ; forgetting to
do the make clean between the parallel and cuda.

All the best,
Jason


>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>



-- 
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 25 2011 - 19:30:02 PDT
Custom Search