[gmx-users] Using the double precision gromacs

[Anton Feenstra]

itai bloch wrote:
> Hi all,
> 
> I have few questions regarding double precision:
> 
> I have Gromacs compiled on my computer in double precision, I would
> like to know if there is a way to make a single precision run without
> having to compile Gromacs again as single precision?
> 
> Is it reasonable to compare results of runs from two different
> starting positions that were done one on single precision and the
> other on double precision? (or that I might have differences due to
> the precision issue)

It is not unreasonable. But, in principle, any two simulations will 
always show differences (unless they are really identical). So, if you 
want to separate differences due to the starting positions from 'other' 
sources of differences (e.g., 'chance', and precision-related things), 
you can compare several (at least 3-5) simulations e.g. with different 
random seeds for starting velocities, and compare (averages from) those.


-- 
Groetjes,

Anton

* NOTE: New Phone & Fax numbers (below) *

  _____________ _______________________________________________________
|             |                                                       |
|  _   _  ___,| K. Anton Feenstra                                     |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|(   |   )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands   |
| \_/ \_/ | | | Tel: +31 20 59 87608 - Fax: +31 20 59 87610           |
|             | Feenstra at few.vu.nl - www.few.vu.nl/~feenstra/         |
|             | "If You See Me Getting High, Knock Me Down" (RHCP)    |
|_____________|_______________________________________________________|