Hi,
I think Xmin error -4 means divide by zero error, but unfortunately
that doesn't tell us much. Do the sander test cases pass, particularly
the lmod-related tests (make test.lmod in $AMBERHOME/test)?
One thing you can try is running on a single processor for a smaller
number of steps and see if that works. I don't know if the lmod / xmin
implementation in sander has been well-tested in parallel with
mvapich2.
-Dan
On Tue, Oct 11, 2011 at 2:54 PM, Sarah Edwards <sarah.sarahjedwards.com> wrote:
> Hello all,
> I am trying to implement LMOD minimization in AMBER11 and have run into the
> following issue:
>
> XMIN ERROR: Status is -4
>
> I've searched the archives and can't find out what that particular error
> means.
>
> My input file is as follows:
> # For running multi-threaded AMBER on the cluster, using the scheduler
> sbset amber11 intel11 mvapich2-testing-pbs-icc11
>
> # low mode md
> cat > lmod.in << EOG
> for lyA-p lyT 10-mer: initial minimization prior to MD
> &cntrl
> imin = 1,
> maxcyc = 2000,
> ntb = 0,
> ntmin = 4,
> ntpr = 1,
> igb = 0,
> cut = 12.0,
> &end
> &ewald
> eedmeth = 5
> &end
> &end
> EOG
>
> # low mode md
> mpiexec -np 8 sander.MVAPICH -O \
> -i lmod.in \
> -o 2JOT-3a.lmod.out \
> -p 2JOTa.top \
> -c 2JOTa.crd \
> -r 2JOT-3a.lmod.crd \
> </dev/null
>
> Another question: Is it possible to perform an LMOD minimization with NMR
> restraints turned on, or does that give non-physical results?
>
> Thanks!
> ~Sarah
> _______________________________________________
> 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 Oct 11 2011 - 13:30:03 PDT
