> Dear Jaimin, although I would be very carefull in saying that a top file and an ali file are "alright", the problem might be somewhere else. I recall from you numerous emails that you are running under windows. I did it once. From my experience, this is the time equation I get when running on various platforms: 2.2Ghz linux athlon = 1 powermac G5 = 1.3 windows2000 700Ghz Athlon = 10++ I agree with you that this list is not a complete benchmark, put there are two points: 1. I'm disappointed by the G5's performance, but this is not the problem you are interested in. 2. Windows was slow, i.e. more than 10 times slower than my mac notebook G4 800GHz. Actually, it was so slow, that 10 times does not mean anything at all. In conclusion, if you get stuck when generating the restraints (which is a costly step anyway), it might actually work, but being so slow you'd think it is stuck. At least if stuck does not mean "crashed". Maybe RAM is the problem ? Is it swapping ? I recall that you said it was working on a linux PC, but that you were not very fond of using linux. Maybe you will loose less time learning this very usefull OS rather than fighting against your windows PC ? ;-) I strongly advise you to switch to linux. BR Antoine > > > > Dear MODELLER users, > > Thank you for ur help so far. > One problem still persists. MODELLER hangs at one point after some > time: > > Dynamically allocated memory at amprmcns [B,kB,MB]: 17981749 > 17560. > 301 17.149 > openf5__224_> Open 11 OLD SEQUENTIAL galk.rsr > openf5__224_> Open 11 OLD SEQUENTIAL galk.rsr > > The TOP and ALI files are alright. > Please solve my problem.
There is also the problem that windows text editors appear to put unreadable characters into, particularly, ali files meaning that modeller doeasn't read them in the way you intended. Thus they look alright but are not. This has been reported to this list before. It would also suggest using Linux. Nigel Quoting Antoine Leimgruber <antoine.leimgruber@etu.unil.ch>: > > > > Dear Jaimin, > > although I would be very carefull in saying that a top file and an ali > file are "alright", the problem might be somewhere else. > > I recall from you numerous emails that you are running under windows. I > did it once. From my experience, this is the time equation I get when > running on various platforms: > > 2.2Ghz linux athlon = 1 > powermac G5 = 1.3 > windows2000 700Ghz Athlon = 10++ > > I agree with you that this list is not a complete benchmark, put there > are two points: > > 1. I'm disappointed by the G5's performance, but this is not the > problem you are interested in. > 2. Windows was slow, i.e. more than 10 times slower than my mac > notebook G4 800GHz. Actually, it was so slow, that 10 times does not > mean anything at all. > > In conclusion, if you get stuck when generating the restraints (which > is a costly step anyway), it might actually work, but being so slow > you'd think it is stuck. At least if stuck does not mean "crashed". > Maybe RAM is the problem ? Is it swapping ? > I recall that you said it was working on a linux PC, but that you were > not very fond of using linux. Maybe you will loose less time learning > this very usefull OS rather than fighting against your windows PC ? ;-) > I strongly advise you to switch to linux. > > > BR > Antoine > > > > > > > > > Dear MODELLER users, > > > > Thank you for ur help so far. > > One problem still persists. MODELLER hangs at one point after some > > time: > > > > Dynamically allocated memory at amprmcns [B,kB,MB]: 17981749 > > 17560. > > 301 17.149 > > openf5__224_> Open 11 OLD SEQUENTIAL galk.rsr > > openf5__224_> Open 11 OLD SEQUENTIAL galk.rsr > > > > The TOP and ALI files are alright. > > Please solve my problem. > > _______________________________________________ > modeller_usage mailing list > modeller_usage@salilab.org > http://salilab.org/mailman/listinfo/modeller_usage > Dr. Nigel Lindsey Senior Lecturer in Biomedical Sciences Department of Biomedical Sciences University of Bradford Bradford W.Yorks Tel +44 01274 233567 ------------------------------------------------------------ This mail sent through IMP: http://webmail.brad.ac.uk To report misuse from this email address forward the message and full headers to misuse@bradford.ac.uk ------------------------------------------------------------
On Wed, Jan 21, 2004 at 04:59:54PM +0000, N.J.Lindsey@Bradford.ac.uk wrote: > There is also the problem that windows text editors appear to put unreadable > characters into, particularly, ali files meaning that modeller doeasn't read > them in the way you intended. Thus they look alright but are not. This has > been reported to this list before. It would also suggest using Linux. This sounds like the perennial problem of different line endings on Unix and Windows machines. If there are other unreadable characters being inserted into alignment files, I'd be intrigued to a) see the resulting files and b) know the editor responsible. Ben Webb, Modeller Caretaker -- modeller-care@salilab.org http://www.salilab.org/modeller/ Modeller mailing list: http://salilab.org/mailman/listinfo/modeller_usage
Dear Modellers, I want to read in atomic coordinates and use modeller to perform MD on the system. In general, I can do this. But I also want to restrain the atoms in the file so that it holds its shape to some extent. So I make a copy of the file and read it in as: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% SET ALNFILE = '/Applications/modeller6v2/ALIGN.ali' # alignment filename SET KNOWNS = 'molecule_copy' SET SEQUENCE = 'molecule' # code of the target SET ATOM_FILES_DIRECTORY = '/Applications/modeller6v2/atom_files' READ_MODEL FILE = 'molecule.pdb', GENERATE_METHOD = 'read_xyz' %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% So now I thought that the easiest way to restrain my model to the original shape was to make homology derived restraints using: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% MAKE_RESTRAINTS RESTRAINT_TYPE = 'distance', ADD_RESTRAINTS = on %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% But I am having an awful time, and I think that I am not defining segments correctly (if at all). I get the following error: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% make_re_417_> Restraint type to be calculated: distance make_re_477E> No templates in the alignment. Number of sequences in the alignment: recover____E> ERROR_STATUS >= STOP_ON_ERROR: 1 1 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Do I need to define my template/known structure more than what is shown above? I have looked through 'homcsr' for ideas, but I couldn't figure it out. Any help would be grateful, Michael ------------------------------------------------------------------------ -------------------------- Michael Grabe, Ph.D. Post-doctoral Fellow Howard Hughes Medical Institute University of California, San Francisco 533 Parnassus Ave. San Francisco, CA 94143 mgrabe@itsa.ucsf.edu tel: ++ 415.476.6585 http://itsa.ucsf.edu/~mgrabe
On Wed, Jan 21, 2004 at 02:20:14PM +0100, Antoine Leimgruber wrote: > I recall from you numerous emails that you are running under windows. I > did it once. From my experience, this is the time equation I get when > running on various platforms: > > 2.2Ghz linux athlon = 1 > powermac G5 = 1.3 > windows2000 700Ghz Athlon = 10++ That sounds about right. The 6v2 Windows binary is not particularly optimized, and will run more slowly than the Linux binary on an equivalent system. I have reproduced the OP's problem - Modeller does not 'hang' at the point of reading the restraints, but simply takes a long time. The job completes in 2 minutes on a 3GHz Xeon running Linux, but in 20 minutes on a 1.6GHz Windows box. We are aware of this problem, and expect there to be little or no difference in runtime between equivalent Linux and Windows machines with the next Modeller release. > I recall that you said it was working on a linux PC, but that you were > not very fond of using linux. Maybe you will loose less time learning > this very usefull OS rather than fighting against your windows PC ? ;-) > I strongly advise you to switch to linux. While we expect to continue to provide Windows and Mac Modeller builds, the majority of development here is conducted on Unix machines, and thus they are likely to always be better supported. Ben Webb, Modeller Caretaker -- modeller-care@salilab.org http://www.salilab.org/modeller/ Modeller mailing list: http://salilab.org/mailman/listinfo/modeller_usage
participants (4)
-
Antoine Leimgruber -
Michael Grabe -
Modeller Caretaker -
N.J.Lindsey＠Bradford.ac.uk