Re: AMBER: pmemd 9.0 compilation error

From: Robert Duke <rduke.email.unc.edu>
Date: Mon, 14 Jan 2008 10:19:51 -0500

Looks to me like you have mpich compiled for use with F77, not F90. The
stuff you see not compiling here is from the mpif.h mpich header, not from
pmemd (Argonne being a tipoff, along with the F77 syntax). When you build
mpich, you have to configure it to use the appropriate compiler (ifort, in
F90 mode). I DO actually allow usage of c/f77 mpi libraries, and that is
what I will typically link to because there is no need for all the f90
interfaces stuff, but you have to have the included headers syntax right,
and there may also be issues with naming conventions on external names in
the linkage step if you don't specify the fortran compiler you plan to use
in linking to mpi libraries. Time to go back and look at the MPI
configuration stuff... (there is some info on the amber web site on this
too, as well as a pretty big README in the pmemd directory).
Regards - Bob Duke

----- Original Message -----
From: "Raphael Chaleil" <raphael.chaleil.cancer.org.uk>
To: <amber.scripps.edu>
Sent: Monday, January 14, 2008 9:38 AM
Subject: AMBER: pmemd 9.0 compilation error


> We tried to compile PMEMD on our cluster ia32 with intel compiler 9.0
> and mpich2_1.0-2.
>
> we used the configure to generate the makefile
>
> maestro amber9/src/pmemd > ./configure linux_p4 ifort mpich2 fftw
> Intel ifort compiler found; version information:
> Version 9.0
> Use Intel MKL? (better performance) (answer yes or no):
> no
> Please enter name of directory where fftw headers are installed:
> /farm/bmm/apps/fftw/include
> Please enter name of directory where fftw libraries are installed:
> /farm/bmm/apps/fftw/lib
> File config_data/fft.fftw being used...
> File config_data/linux_p4.ifort being used...
> File config_data/interconnect.mpich2 being used...
> configure assumes mpich2 files are in /farm/bmm/apps/mpich2.
> PMEMD Configurate successfully completed.
>
> and then tried to compile and got the folowing error
>
> maestro amber9/src/pmemd > make install
> cd src && make install
> make[1]: Entering directory `/farm/bmm/apps/amber9/src/pmemd/src'
> /lib/cpp -traditional -P -I/farm/bmm/apps/fftw/include
> -I/farm/bmm/apps/mpich2/include -DFFTW_FFT -DMPI -DSLOW_NONBLOCKING_MPI
> -DDIRFRC_COMTRANS -DDIRFRC_EFS -DFFTLOADBAL_2PROC parallel_dat.fpp
> parallel_dat.f90
> ifort -c -auto -tpp7 -xN -ip -O3 parallel_dat.f90
> fortcom: Error: parallel_dat.f90, line 34: Syntax error, found
> END-OF-STATEMENT when expecting one of: => = . ( : %
> C
> -------^
> fortcom: Error: parallel_dat.f90, line 35: Syntax error, found
> END-OF-STATEMENT when expecting one of: => = . ( : %
> C
> -------^
> fortcom: Error: parallel_dat.f90, line 36: Syntax error, found
> INTEGER_CONSTANT '2001' when expecting one of: => = . ( %
> C (C) 2001 by Argonne National Laboratory.
> -----------^
> fortcom: Error: parallel_dat.f90, line 37: Syntax error, found
> IDENTIFIER 'SEE' when expecting one of: => = . ( : %
> C See COPYRIGHT in top-level directory.
> -------^
> fortcom: Error: parallel_dat.f90, line 38: Syntax error, found
> END-OF-STATEMENT when expecting one of: => = . ( : %
> C
> -------^
> fortcom: Error: parallel_dat.f90, line 39: Syntax error, found
> IDENTIFIER 'DO' when expecting one of: => = . ( :
> C DO NOT EDIT
> -------^
> fortcom: Error: parallel_dat.f90, line 40: Syntax error, found
> IDENTIFIER 'THIS' when expecting one of: => = . ( : %
> C This file created by buildiface
> -------^
> fortcom: Error: parallel_dat.f90, line 41: Syntax error, found
> END-OF-STATEMENT when expecting one of: => = . ( : %
> C
> -------^
> fortcom: Error: parallel_dat.f90, line 35: This statement is positioned
> incorrectly and/or has syntax errors.
> C
> -------^
> fortcom: Error: parallel_dat.f90, line 41: This statement is positioned
> incorrectly and/or has syntax errors.
> C
> -------^
> compilation aborted for parallel_dat.f90 (code 1)
> make[1]: *** [parallel_dat.o] Error 1
> make[1]: Leaving directory `/farm/bmm/apps/amber9/src/pmemd/src'
> make: *** [install] Error 2
>
> Any idea what's going wrong?
>
> Thanks a lot
>
>
>
> --
> Dr Raphael Chaleil
> Senior Scientific Officer
> Biomolecular Modelling Laboratory
> Cancer Reseach UK London Research Institute
> Tel: +44 (0)20 7269 3380
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber.scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
>


-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Jan 16 2008 - 06:07:13 PST
Custom Search