AMBER Archive (2003)

Subject: Re: Bug found in sander - incorrect conversion factor causes randomized velocities to be too cold

From: David A. Case (case_at_scripps.edu)
Date: Sat Jan 18 2003 - 12:53:54 CST


On Sat, Jan 18, 2003, John D. Chodera wrote:

> I have located what appears to be an error in the AMBER7 version of
> sander. In the file newvel.f, line 58, in subroutine RANVEL, the
> following line appears:
>
> BOLTZ = (8.31441D-3/4.814D0)*TEMP0
>
> The conversion factor from joules to calories is incorrect. The line
> should, of course, read:
>
> BOLTZ = (8.31441D-3/4.184D0)*TEMP0
>
> Looks like somebody's brain was going faster than their fingers. :)
>
> This bug would cause the system temperature to be consistently too cold
> when using the VRAND flag to redraw the velocities at regular intervals.
>

Thanks very much!! I had noticed this exact symptom, but had not seen
why. In fact, I've had "check vrand implementation" on my to-do list for
a while, but didn't see this problem. I'll issue a bug patch ASAP.

Note to users: this bug goes back at least to Amber 4.1, so earlier versions
will have to be updated as well. Fortunately (I guess), RANVEL is hardly
ever called in most simulations. Even if you (incorrectly) randomize
velocities at the beginning of a run, the other temperature controllers (which
don't have this problem) will fix things up pretty quickly. And, many
simulations begin at T=0 and never invoke this routine.

Note to Carlos: this is probably the source of the behavior I was describing
to you at the recent meeting.

Thanks again for the report....dave case

-- 

================================================================== David A. Case | e-mail: case_at_scripps.edu Dept. of Molecular Biology, TPC15 | fax: +1-858-784-8896 The Scripps Research Institute | phone: +1-858-784-9768 10550 N. Torrey Pines Rd. | home page: La Jolla CA 92037 USA | http://www.scripps.edu/case ==================================================================