Giulio Rastelli wrote:
>
> Hi all,
>
> I have found many diffs running the tests of sander with
> Amber 6 (SGI 6.5.8). Most of them arise from very different
> dihedral energies. For example, in the dhfr dir:
> < NSTEP = 1 TIME(PS) = 510.051 TEMP(K) = 302.02 PRESS =
> 0.00
> < Etot = -57716.6180 EKtot = 14145.7441 EPtot =
> -71862.3622
> < BOND = 452.1690 ANGLE = 1277.0334 DIHED =
> 968.3542
> ---
> > NSTEP = 1 TIME(PS) = 510.051 TEMP(K) = 302.03 PRESS =
> 0.00
> > Etot = -53336.9879 EKtot = 14146.1709 EPtot =
> -67483.1589
> > BOND = 452.1690 ANGLE = 1277.0334 DIHED =
> 5347.5575
>
> I don't know if this is a known thing (couldn't find it on the
> amber web), but would be willing to understand why.
>
> Thanks
>
> Giulio Rastelli
> Dipartimento di Scienze Farmaceutiche
> Universita di Modena e Reggio Emilia
> Via Campi 183
> 41100 Modena - ITALY
> -------------------------------------
> tel 0039-059-2055145
> fax 0039-059-2055131
> -------------------------------------
Hi Guilio,
I don't remember finding differences at 6.5.8 but I will look into it.
I need a little more information: are you using sander or
sander_classic? what is the version of the compiler? Are you running on
Origin2000? did you compile the mpi version or serial? were you running
in parallel? which MACHINE file did you use?
Nathalie
--
--------------------------------------------------------------------------
Nathalie Godbout, Ph.D. SGI
godbout_at_boston.sgi.com Chemistry and Biology Applications Group
(978)567-2147 One Cabot Road, Hudson
(978)562-4755 fax MA 01749, USA
--------------------------------------------------------------------------
Received on Fri Oct 20 2000 - 06:14:50 PDT
