Hi Jason,
For years on Kraken I have just always configured AMBER with pgf90 and then
just told pmemd to use pgf90 and mpi. Then just edit the config.h file for
pmemd and replace pgf90 with ftn since ftn is just a wrapper for pgf90. Of
course the problem now is that we no longer support Portland group in the
main AMBER configure script. Thus rather than making a specific crayxt5
option it would make more sense to make Portland group option for the
standard amber configure script that uses pgf90. That way netcdf would be
built with pgf90 and under pmemd it is just a renaming of pgf90 to ftn which
we could add as a cray option.
All the best
Ross
> -----Original Message-----
> From: amber-developers-bounces.ambermd.org [mailto:amber-developers-
> bounces.ambermd.org] On Behalf Of Jason Swails
> Sent: Friday, December 11, 2009 3:06 PM
> To: AMBER Developers Mailing List
> Subject: [AMBER-Developers] configure modified for Cray XT5
>
> Hello,
>
> I'm currently building amber11 on a Cray XT5 machine (Kraken), and
> that option is not currently supported in the configure script. The
> compute nodes in these machines use a special, optimized compiler
> wrapper for each of the compilers (cc, CC, and ftn for c, c++, and
> f77/f90, respectively) which need to be used instead of the actual
> compiler names. The main issue with simply changing these in the
> config.h after it's created is that netcdf configure script was run
> with the other compilers, which I recall having problems with. I
> added a -crayxt5 flag to the configure script. I then added a
> conditional statement after the compiler case statement that sets
> cc=cc, cplusplus=CC, and fc=ftn if crayxt5 is turned on. This works
> exactly as it's supposed to and will result in a successful build on
> Cray XT5 systems.
>
> However, there's one caveat... cc, CC, and ftn exist on the system
> regardless of whether the specified compiler exists (i.e. if intel
> compilers don't exist, ./configure -crayxt5 intel still works since
> the wrappers will simply point to whichever compilers are specified in
> the environment). For this reason, I'm hesitant to commit this change
> to CVS until I've heard more opinions (especially from those higher up
> in the food chain :) ).
>
> Thanks!
> Jason
>
> --
> ---------------------------------------
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Graduate Student
> 352-392-4032
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Dec 11 2009 - 17:30:02 PST
