[Pw_forum] About compilation of cvs version

lan haiping lanhaiping at gmail.com
Sun Sep 21 21:36:06 CEST 2008


Dear Developers:

  I just checked out source code via cvs, and found a compilation error due
to
routine ptoolkit.f90 with information below :
"
IF( ( ii - 1 )/nproc + 1 > SIZE(a,1) ) CALL errore( msg, ' ( ii - 1 )/npr
  The upper bound in the last dimension must appear in the reference to the
assumed size array 'a' at (1). "
My compilor is gfortran v4.1.2.

I then examined this routine ,and found subroutine check_sndbuf_index only
called by cyc2blk_*redist in parallel computation. But no mpi macros  are
defined
for this subroutine. I just wonder we should modify this routine by adding
mpi
macro as following :
#if defined (_MPI)
   IF( j  > SIZE(sndbuf,2) ) CALL errore( msg, ' j > SIZE(sndbuf,2) ', ip+1
)
      IF( il > SIZE(sndbuf,1) ) CALL errore( msg, ' il > SIZE(sndbuf,1) ',
ip+1 )
      IF( ( ii - 1 )/nproc + 1 < 1 ) CALL errore( msg, ' ( ii - 1 )/nproc +
1 < 1 ', ip+1 )
      IF( ( ii - 1 )/nproc + 1 > SIZE(a,1) ) CALL errore( msg, ' ( ii - 1
)/nproc + 1 > SIZE(a,1) ', ip+1 )
      IF( jj < 1 ) CALL errore( msg, ' jj < 1 ', ip+1 )
      IF( jj > n ) CALL errore( msg, ' jj > n ', ip+1 )
#endif
      RETURN


I then recompiled it, and found no error occured again during compiling. I
would like
to know whether it is reasonable by adding this mpi macro for ptoolkit
subroutine.

Best,

Hai-Ping

-- 
Hai-Ping Lan
Department of Electronics ,
Peking University , Bejing, 100871
lanhaiping at gmail.com, hplan at pku.edu.cn
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20080922/6b148915/attachment.html>


More information about the users mailing list