Dear All,
I'm seeing the following failures in AT15 with all bug fixes up to 15
applied. I've compiled this under a 64 bit Ubuntu 10.04.3 LTS OS having
installed the following packages as follows:
sudo apt-get install gcc g++ libxt-dev libxext-dev flex gfortran tcsh
libnetcdf-dev patch csh bison xorg-dev libbz2-dev
possible FAILURE: check hpv.top.dif
possible FAILURE: check hpvbox.top.dif
possible FAILURE: check hpv_cap.top.dif
possible FAILURE: check ubqpol.top.dif
possible FAILURE: check hpvbox.top.dif
possible FAILURE: check hpvoct.top.dif
possible FAILURE: check crn.top.dif
possible FAILURE: check nma1.prmtop.dif
possible FAILURE: check nma2.prmtop.dif
possible FAILURE: check FINAL_RESULTS_MMPBSA.dat.dif
possible FAILURE: check lj.bvv.dif
In decomposing these fails into four classes, with the first set
containing these:
hpv.top.dif
hpvbox.top.dif
hpv_cap.top.dif
ubqpol.top.dif
hpvbox.top.dif
hpvoct.top.dif
crn.top.dif
nma1.prmtop.dif
I think these are related to sleap's prmtop generation and an example of
one is:
possible FAILURE: check hpv.top.dif
/home/mjw/code/AMBER/amber11/AmberTools/test/sleap/prmtop
5c5
< 3128 14 1612 1542 3682 2090 6862 5296
0 0
---
> 3128 14 1612 1542 3682 2090 7082 5296
0 0
The second class of fail is this:
possible FAILURE: check nma2.prmtop.dif
/home/mjw/code/AMBER/amber11/AmberTools/test/sleap/14scale
61c61
< 0. 8.00000000E-1 8.00000000E-2 0. 2.50000000
---
> 8.00000000E-1 0. 8.00000000E-2 0. 2.50000000
This I think is very similar to the first class.
The third is this:
possible FAILURE: check FINAL_RESULTS_MMPBSA.dat.dif
/home/mjw/code/AMBER/amber11/AmberTools/test/mmpbsa_py/06_NAB_Nmode
9c9
< Vibrational: 58.9727 6.1650
> Vibrational: 58.7605 3.1948
The forth is just a rounding issue in the last place:
lj.bvv.dif
The first (and second) class fails are serious, sleap is generating the
wrong prmtop file with reference to the saved output. Can anyone comment
on this? I'm trying to ascertain if I'm dealing with a compiler issue or
an outdated *.save sleap comparison file.
Thanks,
Mark
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Thu Sep 01 2011 - 03:30:02 PDT
