From owner-nwchem-users Wed Sep 10 23:09:25 2003 Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.8.8+Sun/8.8.5) id WAA11197 for nwchem-users-outgoing; Wed, 10 Sep 2003 22:54:15 -0700 (PDT) Date: Thu, 11 Sep 2003 00:54:15 -0500 From: Ricky Kendall Subject: Re: NWCHEM binaries and MPI In-reply-to: To: h.fruchtl@fle.fujitsu.com, nwchem@dresworks.com Cc: nwchem-users@emsl.pnl.gov Reply-to: rickyk@ameslab.gov Message-id: <3F600E07.9020904@scl.ameslab.gov> MIME-version: 1.0 Content-type: text/plain; format=flowed; charset=ISO-8859-1 Content-transfer-encoding: 7bit X-Accept-Language: en-us, en User-Agent: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.4) Gecko/20030624 References: Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk Just using the faster MPI won't help much. the GA tools need to be compiled with the faster network as well. h.fruchtl@fle.fujitsu.com wrote: >Yes, you need to compile it differently if you want to use it with MPI. On a >cluster, there's no real advantage to that (at least in my experience), >unless you have a fast MPI library that bypasses TCP/IP and/or uses a faster >network (Myrinet, etc.). > >HTH, > > Herbert >-----Original Message----- >From: nwchem@kressworks.com [mailto:nwchem@kressworks.com] >Sent: 10 September 2003 23:22 >To: nwchem-users@emsl.pnl.gov >Subject: NWCHEM binaries and MPI > >I have downloaded and installed the NWCHEM binaries on my 5 node, 10 >processor >beowulf cluster running RedHat Linux v 7.3. I have run the parallel NWCHEM >using the associated 'parallel' executable and it works fine over my small >cluster. > >The documentation indicates that NWCHEM will also run using MPI and the >'mpirun' command. However, when I try > >mpirun -np 4 -machinefile machines nwchem test_n2.inp > test.out > >I get the following in the output file: > >argument 1 = test_n2.inp > argument 2 = -p4pg > argument 3 = /s1/PI7818 > argument 4 = -p4wd > argument 5 = /s1 > >... > > Job information > --------------- > > hostname = master.bw01.kressworks.com > program = /s1/nwchem > date = Wed Sep 10 22:11:05 2003 > > compiled = Tue May 20 13:23:02 PDT 2003 > source = /home/edo/nwchem-4.5 > nwchem branch = 4.5 > input = water_test.inp > prefix = h2o. > data base = ./h2o.db > status = startup > nproc = 1 > time left = -1s > > >NOTE: the nproc=1, and the -p4pg=/s1/PI7818 > >So, the PROCGRP is apparently being autogenerated by NWCHEM and is called >PI7818. The -machinefile machines specification (which works just fine with > >my other MPI applications) is being ignored. > >The contents of PI7818 are: > >master.bw01.kressworks.com 0 /s1/nwchem >n01.bw01.kressworks.com 1 /s1/nwchem >master.bw01.kressworks.com 1 /s1/nwchem >n01.bw01.kressworks.com 1 /s1/nwchem > >which is consistent with -np 4 but I have NO idea from where NWCHEM is >getting >this info. nproc should be 4, but it is set to 1. > >Do I need to compile NWCHEM on my system to get it to work properly with >MPI? >Where is NWCHEM getting its info for its autogeneration of the PROCGRP file? > >How do I get NWCHEM to recognize the -np 4 directive? > > The manual is a bit sketchy about the details of running NWCHEM with MPI so > >any and all help would be appreciated. > >Jim > > >------------------------------------------------------------------------ > >This e-mail has been scanned by Trend InterScan Software. > >This e-mail (and its attachment(s) if any) is intended for the named >addressee(s) only. It may >contain information which is privileged and confidential within the >meaning of the applicable law. >Unauthorised use, copying or disclosure is strictly prohibited and may >be unlawful. > >If you are not the intended recipient please delete this email and >contact the sender via email return. > >Fujitsu Laboratories of Europe Ltd (FLE) does not accept responsibility >for changes made to this email after >it was sent. The views expressed in this email may not necessarily be >the views held by FLE. > >Unless expressly stated otherwise, this email does not form part of a >legally binding contract >or agreement between the recipient and Fujitsu Laboratories of Europe Ltd (FLE). > >