On Wed, May 21, 2014, Indrajit Deb wrote:
>
> I am facing a peculiar problem in running AMBER (Amber12 and AmberTools13).
> I have complied AMBER on 21/05/2014 in a machine having 64 bit Ubuntu 12.04
> LTS (server edition) with AMD opteron 6200 (16 core). During the NVT
> equilibration process *(tem0=300.0K)* in explicit solvent the temp. reaches
> to ~4000.0K and it stays tiil the NSLIM.
>
> Whereas when I recompiled the same AMBER on the same date in another
> machine having 64bit Ubuntu 12.04 LTS (Desktop version) with Intel Xeon
> E5-2609 (8 core), the same simulation (as mentioned above) goes fine, where
> it satisfies the *temp0*.
> *Output*
Usual advice: run a 20-step MD simulation from the same starting points on
each machine. Set ntpr=1. See if there are any differences.
Of course, this is just a first step. I don't think you have yet said whether
you are doing serial or parallel simulations, or whether you are using sander
or pmemd. Such information would be helpful. (If you have been running
parallel simulations, try the experiment above with a serial calculation.)
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 21 2014 - 08:00:02 PDT
