Re: AMBER: is possible to compile pmemd using scali MPI?

From: Andrea Bortolato <andrea.bortolato.1.unipd.it>
Date: Mon, 23 Apr 2007 14:31:53 +0200

Hi Gustavo,

I'm sorry I answer so late, I completely forgot!

here is the config.h file I used:

MATH_DEFINES =
MATH_LIBS =
IFORT_RPATH = /share/intel/mkl/9.0/lib/em64t
MATH_DEFINES = -DMKL
MATH_LIBS = -L/share/intel/mkl/9.0/lib/em64t/lib/em64t -lmkl_em64t -lguide -lpthread
FFT_INCLUDEDIR = /share/intel/mkl/9.0/include/fftw
FFT_LIBDIR = /share/intel/mkl/9.0/lib/em64t
FFT_DEFINES = -DFFTW_FFT
FFT_INCLUDE = -I$(FFT_INCLUDEDIR)
FFT_LIBS = -L$(FFT_LIBDIR) -lfftw3xf_intel
NETCDF_HOME =
NETCDF_DEFINES =
NETCDF_MOD =
NETCDF_LIBS =
MPI_HOME = /opt/scali
MPI_DEFINES = -DMPI
MPI_INCLUDE = -I$(MPI_HOME)/include64
MPI_LIBDIR = $(MPI_HOME)/lib64
MPI_LIBS = -L$(MPI_LIBDIR) -lmpi -lfmpi
DIRFRC_DEFINES = -DDIRFRC_EFS -DDIRFRC_COMTRANS -DDIRFRC_NOVEC
CPP = /lib/cpp
CPPFLAGS = -traditional -P
F90_DEFINES = -DFFTLOADBAL_2PROC

F90 = ifort
MODULE_SUFFIX = mod
F90FLAGS = -c -auto
F90_OPT_DBG = -g -traceback
F90_OPT_LO = -tpp7 -O0
F90_OPT_MED = -tpp7 -O2
F90_OPT_HI = -tpp7 -xW -ip -O3
F90_OPT_DFLT = $(F90_OPT_HI)

CC = gcc
CFLAGS =

LOAD = ifort
LOADFLAGS =
LOADLIBS = -limf -lsvml -Wl,-rpath=$(IFORT_RPATH)


hope it helps,
best,
Andrea



On Fri, 30 Mar 2007 16:21:39 -0400
Gustavo Seabra <gustavo.seabra.gmail.com> wrote:

> Hi Andrea,
>
> Congrats! You might also consider posting to the list what
> changes/tweaks you needed to make (and of course also pass this info to
> Bob Duke, as he mentioned), so that future users can benefit benefit
> from your experience.
>
> Best regards,
>
> Gustavo.
>
>
> Andrea Bortolato wrote:
> > thanks a lot, I fought a bit, but finally it worked
> > best
> > Andrea
> >
> > On Thu, 29 Mar 2007 12:36:01 -0400
> > "Robert Duke" <rduke.email.unc.edu> wrote:
> >
> >> Undoubtedly, but I have never had access to a machine with scali. You may
> >> want to look at how it is being done in sander and see if that gives you
> >> some clues as to what is required in pmemd. I presume it is mostly a matter
> >> of picking up the headers and libraries correctly. Try doing a fake
> >> configure of mpich (ie., run the pmemd configure command, and act like you
> >> have mpich instead of scali), and then do the following to the config.h for
> >> pmemd (in the pmemd subdir):
> >> 1) remove -DSLOW_NONBLOCKING_MPI from the MPI_DEFINES = line
> >> 2) for the MPI_LIBS line, change -lmpich to -lfmpi -lmpi
> >>
> >> Please let me know how this works.
> >>
> >> Regards - Bob Duke
> >>
> >> ----- Original Message -----
> >> From: "Andrea Bortolato" <andrea.bortolato.1.unipd.it>
> >> To: <amber.scripps.edu>
> >> Sent: Thursday, March 29, 2007 12:17 PM
> >> Subject: AMBER: is possible to compile pmemd using scali MPI?
> >>
> >>
> >>> Dear all,
> >>>
> >>> I installed successfully amber serial and parallel in our cluster with
> >>> scali MPI.
> >>>
> >>> however I cannot compile pmemd: is possible to compile it using scali MPI?
> >>>
> >>> Thank you very much and best regards
> >>> Andrea
> >>>
> >>> ---------------------------------------
>
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>


---------------------------------------
Andrea Bortolato, Ph.D. Student

Molecular Modeling Section
Department of Pharmaceutical Sciences
University of Padova,
Via Marzolo, 5
35131 Padova - ITALY
Phone: +39-049-8275801
Fax: +39-049-8275801
e-mail:andrea.bortolato.1.unipd.it
Webpage URL http://mms.dsfarm.unipd.it
----------------------------------------
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Apr 25 2007 - 06:07:19 PDT
Custom Search