AMBER Archive (2009)

Subject: Re: [AMBER] red problem

From: FyD (fyd_at_q4md-forcefieldtools.org)
Date: Thu Oct 01 2009 - 23:48:57 CDT


Dear James,

> I seem to be having an issue with RED-vIII.2
> RED is only spawning 1 process of PC-GAMESS on my linux box despite having
> edited the $NP variable in the RED pearl script.

I just tested III.2 & PC-GAMESS works in parallel on linux.

                       ---------------------------
                      * Welcome to R.E.D. III.2 *
                          RESP ESP charge Derive
                   http://q4md-forcefieldtools.org/RED/

                          CHARGE TYPE = RESP-A1
                       ---------------------------
        To be used in agreement with the R.E.D.-III tools license.
                       ---------------------------
                   Date: Fri Oct 2 06:19:00 CEST 2009
                   Machine: master0.q4md-forcefieldtools.org
                       ---------------------------
                   Number of cpu(s) used in the QM jobs(s): 2
                       ---------------------------

[cut ...]

                         * Selected QM Software *
                                 PC-GAMESS

[cut ...]

                          * Software checking *
            pcgamess [ OK ]
            resp [ OK ]

[cut ...]

PC-GAMESS output:

  Intel Core2/ Linux PC GAMESS version running under Linux.
  Running on Intel CPU: Brand ID 0, Family 6, Model 23, Stepping 6
  CPU Brand String : Intel(R) Xeon(R) CPU E5420 @ 2.50GHz
  CPU Features : CMOV, MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1,
MWAIT, EM64T
  Data cache size : L1 32 KB, L2 6144 KB, L3 0 KB
  max # of cores/package : 4
  max # of threads/package : 4
  max cache sharing level : 2
  Operating System successfully passed SSE support test.

  PARALLEL VERSION (OpenMPI) RUNNING WITH 2 NODES

However R.E.D. needs the mpirun program to execute pcgamess in
parallel. If mpirun is not found R.E.D. interfaces PC-GAMESS using a
single cpu/core _without_ providing a warning to the user (this is
bad). If mpirun is found, no report to the user is done as well...

Install a mpi implementation as described on the PC-GAMESS web site.

Get III.3 since this lack of information is corrected in the RED .log file -
today or tomorrow ;-)

regards, Francois

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