Loading ...
Sorry, an error occurred while loading the content.

C-Stack overflow

Expand Messages
  • Ben Eisenbraun
    I have a researcher running the generate step of a CNS refinement on a very large data set (~550,000 atoms). The calculations run for about 90 minutes and
    Message 1 of 3 , Mar 6, 2008
    • 0 Attachment
      I have a researcher running the generate step of a CNS refinement on a
      very large data set (~550,000 atoms). The calculations run for about
      90 minutes and then die like this:

      %<CALLST> error encountered: C-Stack overflow, increase CSTKSI

      That setting appears to be in cstack.inc and was initially set to
      2*MAX_ATOM. I increased it to 4 and then 8 and finally to 32, but
      CNS dies at exactly the same number of lines in the output logfile with
      the same error message.

      This is CNS 1.2 built using the latest Intel compilers on a 64-bit linux
      machine.

      The other local changes are MAX_ATOM set to 600000 and MAXB set to
      4*MAX_ATOM.

      Any insight on how to proceed would be much appreciated.

      -ben

      --
      Ben Eisenbraun
      Structural Biology Grid
      http://sbgrid.org/
    • Axel Brunger
      Dear Ben, Turn on the debug level by modifying the line in the script file: evaluate ($log_level=quiet) to evaluate ($log_level=verbose). Other
      Message 2 of 3 , Mar 11, 2008
      • 0 Attachment
        Dear Ben,

        Turn on the debug level by modifying the line in the script file:

         evaluate ($log_level=quiet)

        to

        evaluate ($log_level=verbose).



        Other trouble-shooting suggestions:

        - any strange atom or residue names, i.e., any non-protein or nucleic acid components?

        - you mentioned that this is a recompiled version with larger MAX_ATOM dimensions. 
        Can you try to reproduce the problem with the standard release by breaking up the
        generate job into smaller pieces?  

        - by the way you can generate smaller pieces individually and then merge the resulting
        .mtf files by

        structure @... end
        structure @... end
        structure @... end
        ...

        write structure output=merged.mtf end

        Of course you have to make sure that the individual pieces have unique segid/resids, etc.

        Axel







        On Mar 11, 2008, at 12:02 PM, Ben Eisenbraun wrote:

        Hello Axel,

        At which step exactly does this error message occur?

        How do I tell which step the program is at?  It is outputting messages
        like:

        MAPIC: Atom numbers being modified
        MAPIC: Atom numbers being modified
        %<CALLST> error encountered: C-Stack overflow, increase CSTKSI
           (CNS is in mode: SET ABORT=NORMal END)

        Before the error is encountered.

        Is the "generate.inp" file unmodified?

        I attached a diff of an unmodified generate.inp with the version being
        run.  It looks like the changes are only to valid parameters.

        -ben

        On Mar 6, 2008, at 7:24 AM, Ben Eisenbraun wrote:

        I have a researcher running the generate step of a CNS refinement on a
        very large data set (~550,000 atoms). The calculations run for about
        90 minutes and then die like this:

        %<CALLST> error encountered: C-Stack overflow, increase CSTKSI

        That setting appears to be in cstack.inc and was initially set to
        2*MAX_ATOM. I increased it to 4 and then 8 and finally to 32, but
        CNS dies at exactly the same number of lines in the output logfile  
        with
        the same error message.

        This is CNS 1.2 built using the latest Intel compilers on a 64-bit  
        linux
        machine.

        The other local changes are MAX_ATOM set to 600000 and MAXB set to
        4*MAX_ATOM.


        --
        Ben Eisenbraun
        Structural Biology Grid
        http://sbgrid.org/
        <generate.diff>

        Axel T. Brunger
        Investigator,  Howard Hughes Medical Institute
        Professor of Molecular and Cellular Physiology
        Stanford University

        Email:  brunger@...      
        Phone:  +1 650-736-1031
        Fax:    +1 650-745-1463



      • Ben Eisenbraun
        Hi Axel, ... Thanks for your help. -ben -- Ben Eisenbraun Structural Biology Grid http://sbgrid.org/
        Message 3 of 3 , Mar 20, 2008
        • 0 Attachment
          Hi Axel,

          Just as a quick follow up, this worked for us:

          > - by the way you can generate smaller pieces individually and then
          > merge the resulting
          > .mtf files by
          >
          > structure @... end
          > structure @... end
          > structure @... end
          > ...
          >
          > write structure output=merged.mtf end

          Thanks for your help.

          -ben

          --
          Ben Eisenbraun
          Structural Biology Grid
          http://sbgrid.org/
        Your message has been successfully submitted and would be delivered to recipients shortly.