On Mon, Jul 7, 2014 at 2:21 AM, Alessandro Contini <
alessandro.contini.unimi.it> wrote:
> Dear Amber users and developers,
> I've compiled AmberTools14 with either intel (composer_xe_2013.1.117,
> icc version 13.0.1) and gnu (gcc version 4.6.3) compilers under Ubuntu
> 12.04 (kernel 3.2.0-65-generic, processor Xeon E5506, 2.13GHz).
> Compilation have been done by applying all the available patches. I've
> tested the serial installation and in both version I've experienced a
> failure, although different:
>
> Intel version:
>
> possible FAILURE: check FINAL_RESULTS_MMPBSA_1.dat.dif
> /usr/local/amber14_intel/AmberTools/test/mmpbsa_py/07_Comprehensive
> 39c39
> < Total -13.7036 1.8048 0.8071
> > Total -7.3880 1.7788 0.7955
> 108c108
>
This may not appear benign, but it is. The test results for this
particular test are wildly imprecise given the nature of the test itself.
But such is the tradeoff for having the test run in a reasonable amount of
time.
> ---------------------------------------
>
> gnu version
>
> possible FAILURE: check tlmod.out.dif
> /usr/local/amber14/AmberTools/test/nab
> 1c1
> < Glob. min. E = -128.257 kcal/mol
> > Glob. min. E = -149.169 kcal/mol
> ### Maximum absolute error in matching lines = 2.09e+01 at line 1 field 5
> ### Maximum relative error in matching lines = 1.63e-01 at line 1 field 5
>
This I haven't seen before. What version of the compilers are you using?
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jul 07 2014 - 08:00:02 PDT
