Re: CCL:G98 "writwa" again



See two comments below.
 At 02:28 PM 7/8/99 -0400, Jim Kress wrote:
 >
 >If you are working with G98W on an NT system there are a number of
 >undocumented file size limitations - some caused by NT, others caused by the
 >G98W programmers.
 >
 >For example, with G98W you cannot have a read write file (rwf)  that is
 >larger than 2GB.  You may see one that appears to be larger but G98W will
 >crash if you continue the calculation.
 Does this symptom also show up with G98W running indefinitely (CPU 100%
 use) and no progress ?
 >If you expect your rwf will need to
 >be larger than 2GB you will need to use the multiple rwf option for G98W and
 >specify multiple rwf files that are all LE 2GB (this is an undocumented
 >'feature' of G98W).
 As far as I know G98W does not allow multiple rwf files.
 If you have successfully done this I would appreciate knowing how.
 >Also, the maxdisk parameter cannot exceed 16GB due to
 >32 bit signed integer limitations in G98W.  Also, when you specify multiple
 >rwf files in G98W the rwf file names you use will interact with the other
 >file names you specify for G98W so that the more rwf you specify the fewer
 >characters you can use for other file names.  In the limit of specifying 8
 >rwf files you will have no characters left over for file names of any other
 >files you may want to specify in G98W.  This is another undocumented
 >'feature' that G98W contains.
 >
 >Gaussian refuses to acknowledge these problems exist and they refuse to
 >provide support for their resolution even though I have sent them sample
 >data that conclusively demonstrates all these problems.  This is one reason
 >why I switched to GAMESS.
 >
 >Good luck.
 >
 >
 >Jim
 >
 >Check out my web site  http://www.kressworks.com/
 >It'll blow your mind (politically), stimulate your senses (artistically)
 >and provide scientific insights that boggle the mind!
 >
 >----- Original Message -----
 >From: WeiQuan Tian <wtian (- at -) uoguelph.ca>
 >To: Max Valdez <maxvalde (- at -) ce.fis.unam.mx>
 >Cc: Sigismondo Boschi <abc0 (- at -) sirio.cineca.it>; <CHEMISTRY
 (- at -) ccl.net>
 >Sent: Wednesday, July 07, 1999 2:40 PM
 >Subject: CCL:Erroneous write. write -1 instead of 50250000
 >
 >
 >>
 >>   Hi,
 >>
 >>  Maybe it is not only the problem for disk space. I did a calculation
 >> C26H12O with 6-31G* by BLYP. I have 2 GB disk space for scratch file
 and I
 >> am sure that this calculation doesn't need more than 2GB disk space.
 >> However the job stopped at l1002 for the frequency calculation. So I
 guess
 >> it is due to memory, however, I have 128MB memory in PII350 and I
 included
 >> 120MB for the calculation. I was confused with this error. As in fact I
 >> met the similar problem with UNIX, while I can't remember how I solved
 >> such a problem.
 >>
 >>  Wei Quan Tian
 >>  Univ. of GUelph
 >>  Canada
 >>
 >>
 >> On Wed, 7 Jul 1999, Max Valdez wrote:
 >>
 >> >     Hi everybody
 >> >     That kind of error is due because the disc has no more space
 to
 >write on it.
 >> > Make sure you will have enough space before you send that kind of
 >calculation (or
 >> > when the molecule is too large). i had the same problem a few days
 ago.
 >> >
 >> >     I hope this helps
 >> >     Waves
 >> >     Max
 >> >
 >> > Sigismondo Boschi wrote:
 >> >
 >> > > Hi all,
 >> > >
 >> > > I have got some similar errors with g98 (A6) on an IBM SP
 Power3,
 >> > > compiled with compiler version xlf 5.1. Is it an error due to
 the
 >> > > impossibility of writing some files, or is it a gaussian
 error?
 >> > >
 >> > > I'm recompiling g98 (A7) with the new compiler, in order to
 see it
 >changes
 >> > > something.
 >> > >
 >> > > Thanks in advance.
 >> > >
 >> > > kind regards
 >> > >
 >> > >      Sigismondo Boschi
 >> > >
 >> >
 >>  -------------------------- -------------------------
 --------------------
 >> > > |                          |  http://incm.cineca.it  |
 >|
 >> > > |                           ------------ ------------
 >|
 >> > > |Sigismondo Boschi                       |tel: +39 051
 6171559
 >|
 >> > > |CINECA (High Performance Systems)       |fax: +39 051
 6132198
 >|
 >> > > |via Magnanelli, 6/3                     |e-mail: s.boschi (-
 at -) cineca.it
 >|
 >> > > |40033 Casalecchio di Reno (BO)-ITALY    |
 >|
 >> >
 >>  ---------------------------------------
 ---------------------------------
 >> > >
 >> > > On Wed, 7 Jul 1999, Qadir K. Timerghazin wrote:
 >> > >
 >> > > > Dear CCLers,
 >> > > >
 >> > > > Using G98w I often get such error during hi-level
 calcualtions (MP4,
 >> > > > QCISD(T), etc.):
 >> > > >
 >> > > > Erroneous write. write -1 instead of 50250000.
 >> > > > fd = 2
 >> > > > writwa
 >> > > >
 >> > > > What's is this? How to fix this problem?
 >> > > >
 >> > > >
 >> > > > Best regards,
 >> > > >  Qadir Timerghazin                       mailto:qt (-
 at -) chat.ru
 >> > > >
 >> > >
 >> >
 >> >
 >> >
 >> >
 >> >
 >> >
 >> >
 >>
 >>
 >> CHEMISTRY (- at -) ccl.net -- To Everybody    |   CHEMISTRY-REQUEST (-
 at -) ccl.net -- To
 >Admins
 >> CHEMISTRY-SEARCH (- at -) ccl.net -- archive search    |    Gopher:
 gopher.ccl.net
 >70
 >> Ftp: ftp.ccl.net  |  WWW: http://www.ccl.net/chemistry/   | Jan:
 >jkl (- at -) ccl.net
 >>
 >>
 >>
 >>
 >
 >
 >
 >----- Original Message -----
 >From: Qadir K. Timerghazin <qt (- at -) chat.ru>
 >To: <CHEMISTRY (- at -) ccl.net>
 >Sent: Thursday, July 08, 1999 2:06 PM
 >Subject: CCL:G98 "writwa" again
 >
 >
 >> Dear CCLers,
 >>
 >> Some days ago I wrote:
 >> > Using G98w I often get such error during hi-level calcualtions
 (MP4,
 >> > QCISD(T), etc.):
 >> >
 >> > Erroneous write. write -1 instead of 50250000.
 >> > fd = 2
 >> > writwa
 >> >
 >> > What's is this? How to fix this problem?
 >> Many people have replied me. The common answer is: "this error due
 to
 >> too small disk space for scratch files". I encountered
 >> this error many times before, and solved the problem by enlarging the
 disk
 >> space.
 >> But now, performing G2 calculation of CH3OClO I encountered
 "writwa"
 >> in the beginning of QCI step:
 >>
 >> #P Geom=AllCheck Guess=TCheck QCISD(T,E4T)/6-311G(d,p)
 >>
 >> I had free space large enough (>4 GB) and the rwf file was 1-1.5 GB
 >> after this crash. Then I write Tran=Conven:
 >>
 >> #P G2 maxdisk=4250MB Tran=Conven geom=check
 >>
 >> and QCI step passed without any problems! And the rwf file didn't
 exceed
 >> 2 GB! So, enigmatic "writwa" error doesn't appear only in the
 case of
 >> the lack of disk space.
 >>
 >> Later, this job produced error during
 >> #P Geom=AllCheck Guess=TCheck MP4/6-311G(2df,p)
 >>
 >>
 >> Could anyone explain this?
 >>
 >> Best regards,
 >>  Qadir Timerghazin                         mailto:qt (- at -)
 chat.ru
 >>
 >>
 >>
 >> CHEMISTRY (- at -) ccl.net -- To Everybody    |   CHEMISTRY-REQUEST (-
 at -) ccl.net -- To
 >Admins
 >> CHEMISTRY-SEARCH (- at -) ccl.net -- archive search    |    Gopher:
 gopher.ccl.net
 >70
 >> Ftp: ftp.ccl.net  |  WWW: http://www.ccl.net/chemistry/   | Jan:
 >jkl (- at -) ccl.net
 >>
 >>
 >>
 >>
 >
 >
 >
 >-= This is automatically added to each message by mailing script =-
 >CHEMISTRY (- at -) ccl.net -- To Everybody    |   CHEMISTRY-REQUEST (- at -)
 ccl.net -- To
 Admins
 >MAILSERV (- at -) ccl.net -- HELP CHEMISTRY or HELP SEARCH
 >CHEMISTRY-SEARCH (- at -) ccl.net -- archive search    |    Gopher:
 gopher.ccl.net 70
 >Ftp: ftp.ccl.net  |  WWW: http://www.ccl.net/chemistry/   | Jan: jkl (- at -) ccl.net
 >
 >
 >
 >
 Laurence
 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
 """"""""""""""""
 Laurence Lavelle, Ph.D.
 University of California Los Angeles
 Department of Chemistry & Biochemistry and Molecular Biology Institute
 Laboratory of Structural Biology & Molecular Medicine
 Los Angeles, CA 90095-1570, USA
 Email:LAVELLE (- at -) MBI.UCLA.EDU
 Phone (Office): (310) 825-2083
 Room 3048A Young Hall
 Fax: (310) 206-4038
 Phone (Lab): (310) 206-8270
 Room 269B MBI
 http://www.doe-mbi.ucla.edu/people/lavelle/lavelle.html
 """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
 """"""""""""""""