On Wed, Dec 17, 2008, Ross Walker wrote:
>
> This reminds me of something we should probably discuss at the AMBER
> developers meeting. I propose that we change the way ig works and we should
> try to make sure this gets put into both sander and pmemd. My proposal is
> that +ve values of ig should behave as they do now. Then if one sets a value
> of ig=0 then it uses the time of day in microseconds for the random number
> stream.
This microsecond model is currently triggered by ig=-1. Why should be
introduce backward compatibility with a change. I don't think anyone
needs to use 1 or -1 as a true seed.
> In the case of ig=0 we would need to decide whether or not random numbers
> are synchronized between threads but I vote not to do this since for
> Langevin as far as I can tell the current approach of enforcing the
> synchronization of random numbers is crazy and doesn't help with the actual
> accuracy at all - in fact it allows people to introduce very weird
> correlations.
Just a note: for things like thermodynamic integration, the code
requires synchonization, since you need to be running the same
trajectory for both groups. [The current ig=-1 option does not work for
TI.]
But I second Bob's suggestion that we learn from what others have done
on parallel random number schemes, and try to avoid "mere" continual
tweaking of what we already have.
...dac
Received on Fri Dec 19 2008 - 01:17:17 PST
