RE: error (157) forrtl: severe

Joachim, Any problem of this nature that is repeatable needs to be looked into. The fact that it happens with some ADVAN's and not others could certainly be a symptom of a bug, e.g., access outside an array bound. The recent email from [EMAIL PROTECTED] contains a link to an interesting discussion of memory allocation bugs ("Don't Touch Me There - What error 157 (Access Violation) is trying to tell you"). I am also concerned about a recent thread "How to increase total image size in Fortran?" having to do with the message warning LNK4084: total image size 772845568 exceeds max (268435456); image may not run. I assume that this message did not occur in your case, just the access violation. But I am uneasy - if it runs despite the message, I wonder if is it possible that some memory addresses could be "black holes"? Please send me a control stream and data that give the Access Violation. I will run it here with subscript checking and other diagnostics enabled which might (if I am lucky) help me locate the problem. On Fri, 7 Dec 2007 14:55:18 +0100, [EMAIL PROTECTED] said: > Yet, also consider this. It is not entirely random. Runs which once > worked , always work, Runs which once crashed with the 157 error > always crash even at the same point (it seems to me at least) into the > execution... as if "something" fills up and then overflows. > > Joachim > > > > > Mark Sale - Next Level Solutions <[EMAIL PROTECTED]> > 07.12.2007 14:44 > > To [EMAIL PROTECTED] cc [email protected] Subject > RE: [NMusers] error (157) forrtl: severe > > > > > > > > > Interesting, The only thing I can think of that would be random like > that is that some other program (virus checker? back up?) is accessing > the files. Maybe the slower ADVANS don't mind because they access the > file(s) less often, and haven't happened to try when the other program > is accessing them. Mark > > > Mark Sale MD Next Level Solutions, LLC www.NextLevelSolns.com > 919-846-9185 > > -------- Original Message -------- Subject: RE: [NMusers] error (157) > forrtl: severe From: [EMAIL PROTECTED] Date: Fri, > December 07, 2007 7:56 am To: [EMAIL PROTECTED] > Cc: [email protected] > > > Thank you, Mark for responding. > > Yes, NONMEM is running often just fine - up to a few hundred function > evaluations. But sometimes it crashes even before I see the first > iteration. When it crashes later I see the entire set of FILEs and > executables in the directory. So NONMEM and NMTRAN were working fine. > > But, how do you explain that the same models reformulated with ANDAN6 > run? Well, they take very long... and none of them has run to a > successful minimization yet. > > Joachim > > > > > > Mark Sale - Next Level Solutions <[EMAIL PROTECTED]> > 07.12.2007 13:44 > > To [EMAIL PROTECTED] cc [email protected] Subject > RE: [NMusers] error (157) forrtl: severe > > > > > > > > > > > Joachim, This is probably a file access error, from Fortan/OS, not > from NONMEM. Does NMTRAN run? Does the compiler run? Does nmlink6 > run (do you get link.lnk)? Do you get an executable? Does NONMEM > start then stop? If you get the executable (nonmem.exe or fsubs.exe), > then the error is with file access in the NONMEM itself. Don't know > which file, but if NMTRAN runs it isn't the (NMTRAN) data set, still > could be FDATA, FCON, FSTREAM, FRECORD, any of the FILE? (but probably > not) or an MSF. Fortran should close all the files when it exits, by > any method (e.g., if it crashes). Could conceivable be a library > file from the compiler as well. Mark > > > Mark Sale MD Next Level Solutions, LLC www.NextLevelSolns.com > 919-846-9185 > > -------- Original Message -------- Subject: [NMusers] error (157) > forrtl: severe From: [EMAIL PROTECTED] Date: Fri, > December 07, 2007 6:19 am To: [email protected] > > > A question for FORTRAN experts: > > Fortran Error 157: access violation > > is the error we see when NONMEM crashes on our server installation > with models which otherwise run fine on our local PC installations of > NONMEM. These crashes occur with very simple (ADVAN2) and also with > rather complicated (ADVAN5, ADVAN7) models. They can be avoided when > models are reformulated with ADVAN6 TOL=5. Of course with ADVAN6 all > models run slower and this is not really a solution of our problem. > > We run NONMEM VI without the recent bug fixes in a "naked" fashion on > the server, and with "wings" on the PCs. Our compiler is Compaq Visual > Fortran (latest version). > > We appreciate your thoughts on that. > > __________________________________________ > Joachim GREVEL, Ph.D. MERCK SERONO International S.A. Exploratory > Medicine 1202 Geneva Tel: +41.22.414.4751 Fax: +41.22.414.3059 Email: > [EMAIL PROTECTED] > > This message and any attachment are confidential, may be privileged or > otherwise protected from disclosure and are intended only for use by > the addressee(s) named herein. If you are not the intended recipient, > you must not copy this message or attachment or disclose the contents > to any other person. If you have received this transmission in error, > please notify the sender immediately and delete the message and any > attachment from your system.