From owner-nwchem-users@emsl.pnl.gov Mon Feb 11 04:18:44 2008 Received: from odyssey.emsl.pnl.gov (localhost [127.0.0.1]) by odyssey.emsl.pnl.gov (8.14.1/8.14.1) with ESMTP id m1BCIhLW004007 for ; Mon, 11 Feb 2008 04:18:44 -0800 (PST) Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.14.1/8.14.1/Submit) id m1BCIhLg004006 for nwchem-users-outgoing-0915; Mon, 11 Feb 2008 04:18:43 -0800 (PST) X-Authentication-Warning: odyssey.emsl.pnl.gov: majordom set sender to owner-nwchem-users@emsl.pnl.gov using -f X-Ironport-SG: OK_Domains X-Ironport-SBRS: 4.7 X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AgAAABPMr0ee4wIRk2dsb2JhbACQOQEBAQEHCgkglks X-IronPort-AV: E=Sophos;i="4.25,333,1199692800"; d="scan'208";a="66356569" From: "Txema Mercero (En)" To: users nwchem Subject: [NWCHEM] Problems with parallel PSPW Date: Mon, 11 Feb 2008 13:18:35 +0100 User-Agent: KMail/1.9.6 Organization: UPV/EHU MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200802111318.35599.jm.mercero@ehu.es> X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-3.0 (lgsx17.lg.ehu.es [158.227.2.17]); Mon, 11 Feb 2008 13:18:37 +0100 (MET) Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk Hi there, I am trying to compile nwchem 5.1 in an itanium cluster, with openinfiniband. I get to compile nwchem, and most of the tests are fine. However, I am not able to run IN PARALLEL non of the NWPW tests. I can run them using mpi with one processor, but not with 2 (neither on one node nor on 2.) I do not get any error, when I use 2 CPUS on the same node the calculations hangs just before the "NWPW PSPW Calculation" should appear (see below). And I get: ----- nwchem_mpi: Rank 0:1: MPI_Recv: ibv_poll_cq(): bad status 1 nwchem_mpi: Rank 0:1: MPI_Recv: self cn13 peer cn14 (rank: 0) nwchem_mpi: Rank 0:1: MPI_Recv: error message: local length error(fork() issue?) ---- When I use 2 cpus on 2 different nodes. I am using intel 10.023 compilers (either for CC and FC). I get the same problem with hpmpi and openmpi (for the calculation on one node). This are the variables I have used for the compilation: setenv NWCHEM_TOP /usr/local/nwchem-5.1 setenv LARGE_FILES TRUE setenv NWCHEM_TARGET LINUX64 setenv USE_MPI y setenv USE_MPIF y setenv MPI_LOC /software/openmpi-1.2.5_intel10 setenv MPI_LIB $MPI_LOC/lib setenv MPI_INCLUDE $MPI_LOC/include setenv LIBMPI "-lfmpi -lmpi -lpmpi" setenv BLASOPT "-L/opt/intel/mkl10/lib/64 -lmkl -lmkl_lapack" setenv ARMCI_NETWORK OPENIB Thank you for your help! ============================================================================== internuclear angles ------------------------------------------------------------------------------ center 1 | center 2 | center 3 | degrees ------------------------------------------------------------------------------ 2 Cl | 1 C | 3 H | 108.29 2 Cl | 1 C | 4 H | 108.29 2 Cl | 1 C | 5 H | 108.29 3 H | 1 C | 4 H | 110.62 3 H | 1 C | 5 H | 110.62 4 H | 1 C | 5 H | 110.62 ------------------------------------------------------------------------------ number of included internuclear angles: 6 Here is where it hangs....