From owner-nwchem-users@emsl.pnl.gov Wed Jul 2 07:31:53 2008 Received: from odyssey.emsl.pnl.gov (localhost.pnl.gov [127.0.0.1]) by odyssey.emsl.pnl.gov (8.14.1/8.14.1) with ESMTP id m62EVq4t029417 for ; Wed, 2 Jul 2008 07:31:53 -0700 (PDT) Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.14.1/8.14.1/Submit) id m62EVqhs029416 for nwchem-users-outgoing-0915; Wed, 2 Jul 2008 07:31:52 -0700 (PDT) X-Authentication-Warning: odyssey.emsl.pnl.gov: majordom set sender to owner-nwchem-users@emsl.pnl.gov using -f X-IronPort-AV: E=Sophos;i="4.27,737,1204531200"; d="scan'208";a="78967513" User-Agent: Microsoft-Entourage/11.3.3.061214 Date: Wed, 02 Jul 2008 07:31:46 -0700 Subject: Re: [NWCHEM] NWChem-5.1 and pspw calculations From: "Bylaska, Eric J" To: Bryan Putnam , "nwchem-users@emsl.pnl.gov" Message-ID: Thread-Topic: [NWCHEM] NWChem-5.1 and pspw calculations Thread-Index: AcjcUFrdmaXB2khDEd2uqwAbY5qvZg== In-Reply-To: Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-OriginalArrivalTime: 02 Jul 2008 14:31:48.0155 (UTC) FILETIME=[5C2608B0:01C8DC50] Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk The problem might also be with the integer8 fortran bindings in your mpi library. You might also try setenv USE_MPIF y setenv USE_MPIF4 y On 7/2/08 5:05 AM, "Bryan Putnam" wrote: > On Tue, 1 Jul 2008, Dunyou Wang wrote: > >> Bryan, >> >> Please unset your USE_MPIF variable then recompile the nwpw module. >> > > Dunyou, > > That appears to have fixed the problem. Thanks! > > Bryan > >> Best wishes >> Dunyou Wang >> >> >> >> Bryan Putnam wrote: >>> On Tue, 1 Jul 2008, Dunyou Wang wrote: >>> >>>> What's your environmental setting for building the source ? --Dunyou >>> >>> This is my script for building, and no other NWCHEM-related variables were >>> set in my environment. Let me know if you need the output of "env" also. As >>> you can see, I've tried LINUX, LINUX64, and various version of mpich. >>> >>> Thanks, >>> Bryan >>> >>> >>> #!/bin/ksh >>> #------------------------------------------------------------------------ >>> # Korn shell script to build NWChem-5.1, Bryan Putnam, ITaP, 15-AUG-2006 >>> #------------------------------------------------------------------------ >>> # >>> #PBS -l nodes=1:ppn=1,walltime=20:00:00 # >>> >>> set -a >>> NWCHEM_TARGET=LINUX >>> #NWCHEM_TARGET=LINUX64 >>> #NWCHEM_TARGET=LAPI >>> NWCHEM_TOP=$HOME/nwchem-5.1 >>> LARGE_FILES=TRUE >>> TCGRSH=/usr/bin/ssh >>> #ARMCI_NETWORK=SOCKETS >>> >>> # For MPICH build >>> USE_MPI=y >>> USE_MPIF=y >>> #MPI_LOC="/apps/steele/mpich-1.2.7p1/32/p4-pgi-7.1-6" >>> #MPI_LOC="/apps/steele/mpich-1.2.7p1/64/p4-pgi-6.0-5" >>> #MPI_LOC="/apps/steele/mpich-1.2.7p1/64/p4-intel-10.1.015" >>> #MPI_LOC="/apps/steele/mpich2-1.0.7/64/ssm-intel-10.1.015" >>> MPI_LOC="/apps/steele/mpich-1.2.7p1/32/p4-intel-9.0.025" >>> MPI_LIB=$MPI_LOC/lib >>> MPI_INCLUDE=$MPI_LOC/include >>> LIBMPI="-lfmpich -lmpich -lpmpich" >>> #LIBMPI="-lfmpich -lmpich" >>> set +a >>> >>> cd $NWCHEM_TOP >>> >>> echo "starting run: `date`" 1>&2 >>> >>> #. ${MODULESHOME}/init/sh; module load intel64/10.1.015 >>> . ${MODULESHOME}/init/sh; module load intel32/9.0.025 >>> >>> find . -name *.o -print -exec rm -f '{}' \; >>> find . -name *.a -print -exec rm -f '{}' \; >>> rm -rf $NWCHEM_TOP/$NWCHEM_TARGET/* >>> >>> cd $NWCHEM_TOP/src >>> make realclean >>> make nwchem_config NWCHEM_MODULES=all >>> #make FC=g77 CC=gcc > make_linux.log 2>&1 >>> make FC=ifort CC=icc > make_linux.log 2>&1 >>> #make FC=pgf77 CC=pgcc > make_linux.log 2>&1 >>> >>> cd $NWCHEM_TOP/src/util >>> make version >>> make >>> cd $NWCHEM_TOP/src >>> make link >>> >>> echo "run complete: `date`" 1>&2 >>> >>> >>>> Bryan Putnam wrote: >>>>> Hi, >>>>> >>>>> I'm attempting to run on of the nwchem sample examples in the >>>>> examples/pspw directory using NWChem-5.1, specifically NaCl.nw. Our >>>>> system is a LINUX64 cluster of 8-processor nodes running >>>>> >>>>> Red Hat Enterprise Linux AS release 4 (Nahant Update 6) >>>>> >>>>> I'll get a different error depending on how many nodes/processors I >>>>> select, but if I attempt to run on 2 processors on a single node, I'll >>>>> get the error >>>>> >>>>> 0:0:Pack_init:error pushing stack:: 0 >>>>> p0_25114: p4_error: : 0 >>>>> >>>>> Note that we've built nwchem on top of one of our mpich-1.2.7p1 builds. >>>>> >>>>> This job runs successfully using our build of NWChem-4.7, which was >>>>> built in >>>>> the same manner, on the same OS, same system, etc. Any help would be >>>>> appreciated! >>>>> >>>>> Thanks, >>>>> Bryan >>>>> >>>>> >>>>> -- >>>>> Bryan Putnam >>>>> Rosen Center for Advanced Computing, Purdue University >>>>> Young Hall (Rm. 519) >>>>> 302 Wood Street >>>>> West Lafayette, IN 47907-2108 >>>>> Ph 765-496-8225 Fax 765-494-0566 >>>>> bfp@purdue.edu >>>>> http://www.rcac.purdue.edu >>>>> >>>>> >>>>> >>>> >>> >>> >>> >> >> > >