AMBER Archive (2005)

Subject: AMBER: Unexpected stop of Sander run

From: Grzegorz Jezierski (jezierski_at_theochem.tu-muenchen.de)
Date: Sun May 22 2005 - 14:44:55 CDT


Dear users,
While trying to run Sander in short test simulations, I have encoutered
strange errors which stopped the execution of a program. I used to work
with AMBER package, and I think I checked all the options carefully this
time, so probably the error is non-trivial.
Here I tried several structures (peptides and nucleic acids), also with
and without MPI, with Amber6 and Amber8, and on three different machines
(#1, Opteron cluster; #'s 2 and 3, SGI).

Here is what is happening:
Machine #1 (MPI):
=====
? FORTRAN Runtime Error:
? Variable not present in current namelist
? READ(UNIT=5,...
=====

Machine #2 gives a lot of text (ellipsis for ommitted text):
=====
MPI: On host flugor, Program /scratch/gj/amber6/exe/sander, Rank 0,
Process 595681 received signal SIGTRAP(5)

MPI: --------stack traceback-------
PC: 0x7165080 MPI_SGI_stacktraceback in /usr/lib32/libmpi.so
PC: 0x71655a8 first_arriver_handler in /usr/lib32/libmpi.so
PC: 0x7165858 slave_sig_handler in /usr/lib32/libmpi.so
PC: 0xfafcacc _sigtramp in /usr/lib32/libc.so.1
PC: 0x1006e3e4 pmesh_kspace_get_sizes in /scratch/gj/amber6/exe/sander
PC: 0x1006a258 ewald_mem in /scratch/gj/amber6/exe/sander
PC: 0x100376cc locmem in /scratch/gj/amber6/exe/sander
PC: 0x1002b954 RDPARM in /scratch/gj/amber6/exe/sander
PC: 0x10030600 MDREAD in /scratch/gj/amber6/exe/sander
PC: 0x10008178 SANDER in /scratch/gj/amber6/exe/sander
PC: 0xad39d74 main in /usr/lib32/libftn.so
(...)
MPI: -----stack traceback ends-----
MPI: On host flugor, Program /scratch/gj/amber6/exe/sander, Rank 0,
Process 595681: Dumping core on signal SIGTRAP(5) into directory
/scratch/gj/guanine/test

=====

Machine #3, when run without "mpirun", starts the simulation, starts
writing to the output file and then stops writing the following at the
end of output file:
" rdparm: a parameter array overflowed
            (e.g. the table of dihedral params)"

When MPI was on, the calculation doesn't even start :
=====
MPI: could not run executable
MPI: No details available, no log files found (all_signal.c:203)
=====

Can anyone tell me what is happening? I am not able to even reproduce
the simplest simulation I run several years ago, no matter what machine
I am using.
Thank you.
Grzegorz
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu