[AMBER] pmemd.MPI with a single-entry group file?

From: Kevin Keane <kkeane.sandiego.edu>
Date: Thu, 22 Mar 2018 16:12:25 -0700


I am trying to get Amber to work on our MPI cluster for one of our
researchers, and am running into a problem. Please forgive if I get any
chemical terminology wrong; I'm an IT guy.

I am calling pmemd.MPI with a group file to take advantage of our cluster,
such as this:

/opt/amber/amber16/bin/pmemd.MPI -ng 8 -groupfile

And the groupfile contains eight lines such as this:

-O -i /data/xxx/AMBER/sn/mdin_eq1 -p /data/xxx/AMBER/sn/prmtop.SN -c
/data/xxx/AMBER/sn/sn_25_min.x -ref /data/xxx/AMBER/sn/sn_25_min.x -r
/scratch/xxx/AMBER/9338.mn01/sn_25_eq1.x -o

This works beautifully. Most of the groups I'm processing have eight
elements, but sometimes I have groups with fewer.

When I have a group with only one element, pmemd.MPI fails with the error
message that the file mdin is missing, even though the group file actually
specifies a different file name (-i mdin_eq1). I verified that the mdin_eq1
file exists.

command line:

/opt/amber/amber16/bin/pmemd.MPI -ng 1 -groupfile

Group file:

-O -i /data/xxx/AMBER/sn/mdin_eq1 -p /data/xxx/AMBER/sn/prmtop.SN -c
/data/xxx/AMBER/sn/sn_31_min.x -ref /data/xxx/AMBER/sn/sn_31_min.x -r
/scratch/xxx/AMBER/9341.mn01/sn_31_eq1.x -o


  Unit 5 Error on OPEN:

MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.

  Unit 5 Error on OPEN:

Do I need to provide any other arguments that I overlooked? Or is a group
file with only one entry not supported?

I am of course hoping to avoid having to special-case this scenario.

Kevin Keane | Systems Architect | University of San Diego ITS |
Maher Hall, 192 |5998 Alcalá Park | San Diego, CA 92110-2492 | 619.260.6859
AMBER mailing list
Received on Thu Mar 22 2018 - 16:30:02 PDT
Custom Search