From owner-nwchem-users@emsl.pnl.gov Tue Jul 19 13:33:23 2005 Received: from odyssey.emsl.pnl.gov (localhost [127.0.0.1]) by odyssey.emsl.pnl.gov (8.12.10/8.12.10) with ESMTP id j6JKXNDB011576 for ; Tue, 19 Jul 2005 13:33:23 -0700 (PDT) Received: (from majordom@localhost) by odyssey.emsl.pnl.gov (8.12.10/8.12.10/Submit) id j6JKXMEY011575 for nwchem-users-outgoing; Tue, 19 Jul 2005 13:33:22 -0700 (PDT) Date: Tue, 19 Jul 2005 14:33:19 -0600 From: "David V. Dearden" Subject: Re: failure of large analytical vibrational calculations In-reply-to: <42DD3360.9030402@pnl.gov> To: Edoardo Apra` Cc: nwchem-users@emsl.pnl.gov Message-id: MIME-version: 1.0 (Apple Message framework v733) X-Mailer: Apple Mail (2.733) Content-type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Content-transfer-encoding: 7bit X-Scanned-By: MIMEDefang 2.42 References: <1F02B02B-142B-4318-ADBB-F1FA38533C12@byu.edu> <42DD3360.9030402@pnl.gov> Sender: owner-nwchem-users@emsl.pnl.gov Precedence: bulk On Jul 19, 2005, at 11:07 AM, Edoardo Apra` wrote: > David > the error on your output file > >0:0:dai_zero_eof: write error :: 0 > is very likely due to I/O issues, for example running out of disk > space. Can you check if something like this occurred on your IBM SP? > > Edo > > David V. Dearden wrote: > > >> I am attempting B3LYP calculations of analytical vibrational >> spectra for molecules with about 100 atoms on an IBM SP system. >> All are failing at the same point in the calculation, with the >> following error: >> Multipole analysis of the density >> --------------------------------- >> >> L x y z total alpha beta >> nuclear >> - - - - ----- ----- ---- >> ------- >> 0 0 0 0 2.000000 -233.000000 -233.000000 >> 468.000000 >> >> 1 1 0 0 0.000704 0.000352 0.000352 >> 0.000000 >> 1 0 1 0 0.000234 0.000117 0.000117 >> 0.000000 >> 1 0 0 1 0.000178 0.000089 0.000089 >> 0.000000 >> >> 2 2 0 0 -161.435881 -5549.061567 -5549.061567 >> 10936.687254 >> 2 1 1 0 -0.002037 0.012053 0.012053 >> -0.026142 >> 2 1 0 1 -0.025604 0.060029 0.060029 >> -0.145661 >> 2 0 2 0 -161.408811 -5551.622089 -5551.622089 >> 10941.835367 >> 2 0 1 1 0.194144 -0.431334 -0.431334 >> 1.056811 >> 2 0 0 2 -239.789153 -3743.131075 -3743.131075 >> 7246.472998 >> >> 0:0:dai_zero_eof: write error :: 0 >> 0:ARMCI aborting 0(0) >> >> It looks like I'm getting a write error at the point the file for >> the Hessian should be created. I do get past this point when >> running the same jobs (slowly) on a Linux cluster, and smaller >> jobs do work on the SP. Does anyone know of a workaround to get >> these bigger jobs to go on the SP? >> >> Thanks, >> David >> Edo, I checked disk space and the drive where I was running had about 150 GB free. I had not realized there was that little capacity. I changed to run from the scratch drive, which had about 3.2 TB free, but still had the failure in the exact same place. Is there a simple way to determine how much storage will be needed for a vibrational job? Thanks, David ******************************************* David V. Dearden Professor of Chemistry Department of Chemistry & Biochemistry C100 Benson Science Building Brigham Young University Provo, Utah 84602-5700 USA (801)422-2355 david_dearden@byu.edu