This is my bad from when I fixed the problem with the average command when
the number of frames could not be calcd in advance. I was updating maxFrames
with the value of processed, which of course with ptraj.MPI for each
separate thread is going to be your actual frames divided by the number of
threads. I've now modified it so that the maxFrames variable is only updated
if it could not be pre-calculated (which automatically means it wont happen
for parallel runs since in order to run parallel you need to calc the number
of threads). Attached is a patch which should do the trick.
-Dan
On Sun, Apr 18, 2010 at 7:34 PM, Scott Brozell <sbrozell.rci.rutgers.edu>wrote:
> Hi,
>
> New to this RC are 5 diffs in test_at_parallel for ptraj_comprehensive:
>
> SRB: pgi 9.0-4; x86_64; glenn.osc.edu; RC 18apr10
> possible FAILURE: check dist_end_to_end.list.dif
> /tmp/pbstmp.3066906/amber11/AmberTools/test/ptraj_comprehensive
> 51,100d50
> < 51.00 16.726488
> < 52.00 16.099388
> < 53.00 16.908603
> < 54.00 17.451692
> ...
>
> 2010-04-18_17-42-21.diff is attached.
>
> scott
>
> On Sun, Apr 18, 2010 at 05:11:19AM -0400, David A. Case wrote:
> > My current guess is that the Amber release will be delayed some...I am
> > trapped in Munich, with no clear idea of when I will be able to return to
> > the States.
> >
> > There were unfortunately a lot of glitches in the last release, because I
> > let new tests be added, which was probably a mistake. There is a new RC:
> >
> > http://ambermd.org/downloads/AmberTools.18apr10.tar.bz2
> > http://ambermd.org/downloads/Amber11.18apr10.tar.bz2
> >
> > If you didn't have problems with the 14apr10 RC, you don't necessarily
> need
> > to test this one again...I appreciate the hard work people have put in,
> and
> > don't want to add to your burdens.
> >
> > If you need to make final tweaks to the manuals, please create diffs
> against
> > these files:
> >
> > http://ambermd.org.downloads/AmberDocs.18apr10.tar.bz2
>
> typo correction:
> http://ambermd.org/downloads/AmberDocs.18apr10.tar.bz2
>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
>
--
-------------------------
Daniel R. Roe
Postdoctoral Associate
SAS - Chemistry & Chemical Biology
610 Taylor Road
Piscataway, NJ 08854
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Mon Apr 19 2010 - 06:00:08 PDT
