AMBER Archive (2007)

Subject: AMBER: replica exchange out files continually overwritten

From: David Mobley (dmobley_at_gmail.com)
Date: Fri Dec 14 2007 - 11:02:41 CST

Dear all,

I've used AMBER 9 quite a bit, but am just trying to run my first
replica exchange simulations. I appear to have MPI working properly
and the simulations are running just fine. This is on a xeon cluster.

However, all of my output files for individual replicas (with the
exception of the trajectories) are apparently being overwritten every
exchange attempt. I briefly see contents in them in between exchange
attempts, for example:

-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run0.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run10.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run11.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run12.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run13.out
-rw-r--r-- 1 dmobley dmobley 7028 Dec 14 08:58 run14.out
-rw-r--r-- 1 dmobley dmobley 6830 Dec 14 08:58 run15.out
-rw-r--r-- 1 dmobley dmobley 6830 Dec 14 08:58 run1.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run2.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run3.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run4.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run5.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run6.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run7.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:58 run8.out
-rw-r--r-- 1 dmobley dmobley 7028 Dec 14 08:58 run9.out

But then I see:
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run0.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run10.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run11.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run12.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run13.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run14.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run15.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run1.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run2.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run3.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run4.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run5.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run6.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run7.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run8.out
-rw-r--r-- 1 dmobley dmobley 1 Dec 14 08:59 run9.out

Energy files appear to only have one block of energies ever. However,
my rem.log file fills up just fine, as do my trajectory files.

Here's a sample run input file for one of the replicas:
trajectory segment
 &cntrl
 imin = 0,
 nstlim = 500,
 numexchg = 5000,
 nscm=100,
 igb = 5, gbsa = 1,
 cut=16.0,
 tempi = 300.00,
 temp0 = 300.00,
 ntt = 2,
 vrand = 250,
 ntpr=5000,
 vlimit = 0.000,
 ntc=2, ntf = 2, tol = 1.0d-8
 irest = 1,
 ntx = 5,
 dt = 0.002, nrespa = 2,
 ntb = 0, ntave = 0,
 ioutfm = 0, ntwx = 500, ntwe = 500,
 ntr = 0,
 repcrd = 0
 &end
END
END

And here's my groupfile:
-A -i run0.in -o run0.out -c incrd0.crd -x run0.crd -r run0.rst -p
peptide_probes.prmtop -inf run0.info -e run0.en -rem 1
-A -i run1.in -o run1.out -c incrd1.crd -x run1.crd -r run1.rst -p
peptide_probes.prmtop -inf run1.info -e run1.en -rem 1
-A -i run2.in -o run2.out -c incrd2.crd -x run2.crd -r run2.rst -p
peptide_probes.prmtop -inf run2.info -e run2.en -rem 1
-A -i run3.in -o run3.out -c incrd3.crd -x run3.crd -r run3.rst -p
peptide_probes.prmtop -inf run3.info -e run3.en -rem 1
-A -i run4.in -o run4.out -c incrd4.crd -x run4.crd -r run4.rst -p
peptide_probes.prmtop -inf run4.info -e run4.en -rem 1
-A -i run5.in -o run5.out -c incrd5.crd -x run5.crd -r run5.rst -p
peptide_probes.prmtop -inf run5.info -e run5.en -rem 1
-A -i run6.in -o run6.out -c incrd6.crd -x run6.crd -r run6.rst -p
peptide_probes.prmtop -inf run6.info -e run6.en -rem 1
-A -i run7.in -o run7.out -c incrd7.crd -x run7.crd -r run7.rst -p
peptide_probes.prmtop -inf run7.info -e run7.en -rem 1
-A -i run8.in -o run8.out -c incrd8.crd -x run8.crd -r run8.rst -p
peptide_probes.prmtop -inf run8.info -e run8.en -rem 1
-A -i run9.in -o run9.out -c incrd9.crd -x run9.crd -r run9.rst -p
peptide_probes.prmtop -inf run9.info -e run9.en -rem 1
-A -i run10.in -o run10.out -c incrd10.crd -x run10.crd -r run10.rst
-p peptide_probes.prmtop -inf run10.info -e run10.en -rem 1
-A -i run11.in -o run11.out -c incrd11.crd -x run11.crd -r run11.rst
-p peptide_probes.prmtop -inf run11.info -e run11.en -rem 1
-A -i run12.in -o run12.out -c incrd12.crd -x run12.crd -r run12.rst
-p peptide_probes.prmtop -inf run12.info -e run12.en -rem 1
-A -i run13.in -o run13.out -c incrd13.crd -x run13.crd -r run13.rst
-p peptide_probes.prmtop -inf run13.info -e run13.en -rem 1
-A -i run14.in -o run14.out -c incrd14.crd -x run14.crd -r run14.rst
-p peptide_probes.prmtop -inf run14.info -e run14.en -rem 1
-A -i run15.in -o run15.out -c incrd15.crd -x run15.crd -r run15.rst
-p peptide_probes.prmtop -inf run15.info -e run15.en -rem 1

I've also tried this with the -O option rather than -A (which is what
I was using initially) and see apparently the same thing, which is
consistent with what the manual says (this should only affect what
happens to *existing* files).

Have I done something wrong? Why are my output files being overwritten
every exchange rather than appended to?

Thanks!
David
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu