> Just to be sure, in your case, the hang happens *only* in the parallel
> version, right? Do all serial tests pass just fine?
Yes. They aren't "fine", exactly, but they aren't awful:
=====================================================
Finished serial test suite for Amber 11 at Tue May 11 15:17:06 EDT 2010.
make[1]: Leaving directory `/usr/local/amber11_20100507/test'
394 file comparisons passed
13 file comparisons failed
1 test experienced an error
=====================================================
I'm having trouble figuring out which test experienced an error. I'm guessing it's the DFTB set that all say "X not found - Skipping Test"
The failures are all from the ncsu set:
$ grep -A2 FAILURE 2010-05-11_14-17-27.diff | grep amber11_20100507
/usr/local/amber11_20100507/test/ncsu/abmd_ANALYSIS
/usr/local/amber11_20100507/test/ncsu/abmd_ANALYSIS
/usr/local/amber11_20100507/test/ncsu/abmd_FLOODING
/usr/local/amber11_20100507/test/ncsu/abmd_FLOODING
/usr/local/amber11_20100507/test/ncsu/abmd_FLOODING
/usr/local/amber11_20100507/test/ncsu/abmd_UMBRELLA
/usr/local/amber11_20100507/test/ncsu/abmd_UMBRELLA
/usr/local/amber11_20100507/test/ncsu/smd
/usr/local/amber11_20100507/test/ncsu/smd
/usr/local/amber11_20100507/test/ncsu/pmd
/usr/local/amber11_20100507/test/ncsu/pmd
/usr/local/amber11_20100507/test/ncsu/smd2
/usr/local/amber11_20100507/test/ncsu/smd2
> On Tue, May 11, 2010 at 5:41 PM, Lachele Foley <lfoley.ccrc.uga.edu> wrote:
> > Yes, using gcc44/g++44/gfortran44, etc., all 4.4.0, we still get the hang
> on the same neb test.
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Wed May 12 2010 - 09:00:03 PDT
