Hi,
This is a strange error, since both the trajectory frame size and the
size of the coordinate buffers are calculated from the number of atoms
in the input topology file. There shouldn't be an issue unless you
physically run out of memory. Can you post your ptraj input file as
well as the complete output from ptraj (stdin and stderr), along with
what AT1.5 bugfixes you have applied? Also, have you tried the same
input with cpptraj? Does a similar issue occur?
Thanks!
-Dan
On Tue, Aug 23, 2011 at 4:10 PM, Ben Gamari <bgamari.physics.umass.edu> wrote:
> While running ptraj (AmberTools 1.5) on simulation data produced by
> Amber 11 recently we found that it would reliably crash on one
> particular .mdcrd file with glibc throwing an "invalid pointer" error in
> free().
>
> After running ptraj in valgrind, it appears that bufferToXYZ() is
> trashing some memory out of its buffer[1]. Any ideas what might be
> causing this?
>
> The system is x86_64 running Ubuntu 11.04 with gcc 4.5.2.
>
> Cheers,
>
> - Ben
>
>
> [1] Valgrind errors,
>
> ==10858== Invalid write of size 8
> ==10858== at 0x42C8B8: bufferToXYZ (trajectory.c:2471)
> ==10858== by 0x44CC69: ptrajProcessInputCoordinates (ptraj.c:4048)
> ==10858== by 0x453CE9: ptrajProcessTrajectoryFiles (ptraj.c:6954)
> ==10858== by 0x454CFA: ptraj (ptraj.c:7401)
> ==10858== by 0x439344: interface (interface.c:92)
> ==10858== by 0x40287F: main (main.c:218)
> ==10858== Address 0x65fc898 is 0 bytes after a block of size 619,688 alloc'd
> ==10858== at 0x4C28FAC: malloc (vg_replace_malloc.c:236)
> ==10858== by 0x422E01: safe_malloc (utility.c:172)
> ==10858== by 0x454C46: ptraj (ptraj.c:7391)
> ==10858== by 0x439344: interface (interface.c:92)
> ==10858== by 0x40287F: main (main.c:218)
> ==10858==
> ==10858== Invalid write of size 8
> ==10858== at 0x42C8E7: bufferToXYZ (trajectory.c:2474)
> ==10858== by 0x44CC69: ptrajProcessInputCoordinates (ptraj.c:4048)
> ==10858== by 0x453CE9: ptrajProcessTrajectoryFiles (ptraj.c:6954)
> ==10858== by 0x454CFA: ptraj (ptraj.c:7401)
> ==10858== by 0x439344: interface (interface.c:92)
> ==10858== by 0x40287F: main (main.c:218)
> ==10858== Address 0x94257f8 is 0 bytes after a block of size 619,688 alloc'd
> ==10858== at 0x4C28FAC: malloc (vg_replace_malloc.c:236)
> ==10858== by 0x422E01: safe_malloc (utility.c:172)
> ==10858== by 0x454C61: ptraj (ptraj.c:7392)
> ==10858== by 0x439344: interface (interface.c:92)
> ==10858== by 0x40287F: main (main.c:218)
> ==10858==
> ==10858== Invalid write of size 8
> ==10858== at 0x42C916: bufferToXYZ (trajectory.c:2477)
> ==10858== by 0x44CC69: ptrajProcessInputCoordinates (ptraj.c:4048)
> ==10858== by 0x453CE9: ptrajProcessTrajectoryFiles (ptraj.c:6954)
> ==10858== by 0x454CFA: ptraj (ptraj.c:7401)
> ==10858== by 0x439344: interface (interface.c:92)
> ==10858== by 0x40287F: main (main.c:218)
> ==10858== Address 0x6693d88 is 0 bytes after a block of size 619,688 alloc'd
> ==10858== at 0x4C28FAC: malloc (vg_replace_malloc.c:236)
> ==10858== by 0x422E01: safe_malloc (utility.c:172)
> ==10858== by 0x454C7C: ptraj (ptraj.c:7393)
> ==10858== by 0x439344: interface (interface.c:92)
> ==10858== by 0x40287F: main (main.c:218)
> ==10858==
>
> _______________________________________________
> 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 Tue Aug 23 2011 - 15:00:06 PDT
