Dear all,
having recently installed the latest versions of AmberTools12, including
all auto-updates as of June 6th 2012, locally on a Mac and on our linux
farm I discovered ptraj behaves differently on both systems when trying
to compute a time auto-correlation function:
on my Mac, the following script would run through generating the
expected output,
> ptraj my.psf my.script
with my.script as
trajin my.dcd
vector v0 .1 corr .2 order 2
analyze timecorr vec1 v0 tstep 1.0 tcorr 1000.0 out my.out
while under linux it failed with the following and at that time cryptic
warning,
WARNING in ptraj(), analyze timecorr: different attributes for vec1 and
vec2, ignoring command
obviously, this caused me some major headache since the input files were
100% identical, and also the ptraj versions the same, freshly compiled
(Version: "AMBER 12.0 integrated" (4/2012)).
Luckily the advise given at
http://archive.ambermd.org/201107/0355.html
was hinting at how to fix the problem: the linux version was expecting a
"cross-correlation" kind of definition of the auto-correlation I wanted
to compute,
analyze timecorr vec1 v0 vec2 v0 tstep 1.0 tcorr 25000.0 dplr out my.out
instead of
analyze timecorr vec1 v0 tstep 1.0 tcorr 25000.0 dplr out my.out
gave me the desired output, while the Mac version seemed to be less
picky about the syntax.
This looks a bit suspicious to me ... so I hope this report might be of
value to some.
cheers
Ben
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jun 07 2012 - 02:30:02 PDT
