Re: [AMBER] Sander Error

From: Ali Faruqi <alifar76.gmail.com>
Date: Sat, 25 Jun 2011 13:59:22 +0500

The error message I am getting is generalized irrespective of the system I
am trying to simulate. The simulations had been running smoothly since quite
sometime. Then, all of a sudden, this error popped up. At that time, a
protein was being simulated.
To ensure that there were no errors in my .prmtop, .in and .rst input files
of the protein that was being simulated and to verify if the problem is
indeed with sander, I downloaded the standard input files provided by Ross
Walker's tutorial and issued the standard command provided with the
tutorial. The command is as follows:

$AMBERHOME/exe/sander -O -i polyAT_vac_md1_12Acut.in -o
polyAT_vac_md1_12Acut.out -c polyAT_vac_init_min.rst -p polyAT_vac.prmtop
-r polyAT_vac_md1_12Acut.rst -x polyAT_vac_md1_12Acut.mdcrd

Again, the same error occurred. I tried it with different input files of the
previously successful simulation runs, and ended getting the same error.

When I go into the test directory and issue the 'make test' command, I get
different errors at different times. Sometimes I get the following error:

cd cytosine && ./Run.cytosine
At line 310 of file _mdwrit.f
Fortran runtime error: End of file
  ./Run.cytosine: Program error
make: *** [test.sander.BASIC] Error 1

Sometimes it is the same error:

cd nonper && ./Run.nonper
At line 107 of file _dynlib.f
Fortran runtime error: End of file
  ./Run.nonper: Program error
make: *** [test.sander.BASIC] Error 1

Sometimes it is:

cd dmp && ./Run.dmp
At line 705 of file _mdread.f
Fortran runtime error: End of file
  ./Run.dmp: Program error
make: *** [test.sander.BASIC.nopar] Error 1

There is no fixed pattern that I can discern. Line 107 of the dynlib.f file
provides the standard code and there apparently seems nothing wrong with it.


do j = 1,nn (Line 107)
         aamass = amass(iat+j)
         xcm(1) = xcm(1)+xr(j3+1)*aamass
         xcm(2) = xcm(2)+xr(j3+2)*aamass
         xcm(3) = xcm(3)+xr(j3+3)*aamass

         ! each processor knows all coordinates, but only some velocities:
         if (iat+j >= istart .and. iat+j <= iend) then
            vcm(1) = vcm(1)+v(j3+1)*aamass
            vcm(2) = vcm(2)+v(j3+2)*aamass
            vcm(3) = vcm(3)+v(j3+3)*aamass
         end if

         j3 = j3+3
      end do

What's strange is the fact that AMBER was working quite normally. This
problem occurred all of a sudden, without any apparent interference. I would
really appreciate if something can be done about it. Thanks. Regards.

On Fri, Jun 24, 2011 at 8:01 PM, David A. Case <case.biomaps.rutgers.edu>wrote:

> On Fri, Jun 24, 2011, Ali Faruqi wrote:
>
> > Actually, it holds true with every file I use.
> > > > I am getting the following error when I run sander within a minute's
> > > time:
> > > >
> > > > At line 107 of file _dynlib.f
> > > > Fortran runtime error: End of file
>
> Do all the test cases fail as well? If so, there is something seriously
> wrong
> with your installation (but not possible to figure out without more
> information.)
>
> If (most of) the test cases work, try to see what is different about your
> files and those from the test cases. Also, just saying you downloaded some
> files from the web page, without indicating exactly what commands you used,
> makes it impossible for others to spot what the problem would be.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Jun 25 2011 - 02:30:03 PDT
Custom Search