Dear Amber community!
I've got failure when using SCC-DFTB method by Seabra et al. in QM/MM
simulation in Amber10.
The task crashes at the begining of parallel calculation with the
following error in output:
forrtl: severe (24): end-of-file during read, unit 3, file
/home/nilov/amber/amber10/dat/slko/H-C.skf
Image PC Routine Line
Source
sander.MPI 0000000000A1D8DD Unknown Unknown Unknown
sander.MPI 0000000000A1C3E5 Unknown Unknown Unknown
sander.MPI 00000000009B9FE9 Unknown Unknown Unknown
sander.MPI 00000000009599CF Unknown Unknown Unknown
sander.MPI 0000000000959202 Unknown Unknown Unknown
sander.MPI 0000000000982D8A Unknown Unknown Unknown
sander.MPI 0000000000700AB1 Unknown Unknown Unknown
sander.MPI 000000000070306B Unknown Unknown Unknown
sander.MPI 00000000006B367B Unknown Unknown Unknown
sander.MPI 0000000000663907 Unknown Unknown Unknown
sander.MPI 00000000007D32D9 Unknown Unknown Unknown
sander.MPI 00000000005039D2 Unknown Unknown Unknown
sander.MPI 00000000004C93B4 Unknown Unknown Unknown
sander.MPI 00000000004C09A1 Unknown Unknown Unknown
sander.MPI 000000000041E2DC Unknown Unknown Unknown
libc.so.6 00007FF1B5130CF4 Unknown Unknown Unknown
sander.MPI 000000000041E1E9 Unknown Unknown Unknown
MPI process terminated unexpectedly
Exit code -5 signaled from node-63-07
Killing remote processes...forrtl: error (69): process interrupted (SIGINT)
..an so on.
At the same time, I can successfully use other hamiltonians such as AM1 and PM3.
Thus, now I'm not sure that SCC-DFTB has stable parallel
implementation in Amber10.
Any advice will be appreciated!
I've attached input and output files.
Supercomputer system is based on "2 x Intel Xeon E5472 3.0 GHz" nodes
and InfiniBand.
--
Dmitry Nilov,
Lomonosov Moscow State University
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Dec 17 2010 - 03:30:03 PST
