On Sat, Jan 28, 2017, James Kress wrote:
> I tried openMPI 1.6.5 and ran the tests, again. They ran but with one error
> that I do not understand.
>
> cd nfe && ./run-pmemd-parallel.sh
> /bin/sh: ./run-pmemd-parallel.sh: Permission denied
This is our fault, with incorrect permissions. It will be fixed in the next
release, and is not a cause for concern.
Note however, the the "noboxinfo" test you reported earlier is only run
for sander, not for pmemd. The log file you attached was for pmemd. You
might look at the AmberTools test logfile to search for "noboxinfo", to see if
the test indeed PASSED with openMPI 1.6.5. (I'm guessing that it did, since
otherwise you would have seen the hang reported earlier; but it's probably
best to make sure.)
update: I can confirm the report of a hang when using openmpi-1.10.5; the
test passes with the current openmpi-2.0.1. I haven't been able to see
anything suspicious about the code involved (although others could look).
But given that both earlier and later versions of openmpi work as expected,
I think for the moment we will just recommend not using 1.10.
...regards...dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sat Jan 28 2017 - 18:00:02 PST
