[gmx-users] speed issue, gmx runtime estimate = f(t)

[Justin A. Lemkul]

Michael Brunsteiner wrote:
> Dear all, 
> 
> I am trying to optimize runtimes/speed for a series of
> calculations i plan to do with a rather complex system.
> When looking at runtimes I observed one peculiar issue:
> the estimated runtime written to stderr by mdrun with the "-v"
> option keepsgrowing ... in my experience this estimate normally
> converges to a fairly accurate value within a few hundred
> steps. not so here ..
> for example, starting a job at 12:14 gromacs writes to stderr:
> 
> step 100, will finish Wed Apr  6 12:40:35 2011vol 0.97  imb F  1% 
> step 200, will finish Wed Apr  6 12:39:42 2011vol 0.95  imb F  2%
> step 300, will finish Wed Apr  6 12:39:47 2011vol 0.95  imb F  1%
> step 400, will finish Wed Apr  6 12:39:52 2011vol 0.96  imb F  1%
> ...
> step 38400, will finish Wed Apr  6 14:09:11 2011vol 0.91  imb F 29% 
> 

The decrease in DD cell volume coupled with the large increase in force 
imbalance suggests that your system is undergoing changes that cause the DD 
algorithm to slow down.

> so the estimated run time grows steadily from 25 to 110 minutes 
> within about one hour...

This is likely due to the force imbalance.

<snip>

> energygrps           = SOL CC CF
> energygrp_table      = CF CF  CF SOL  CF CC

Tabulated potentials also run quite a bit slower than normal functions.  In 
conjunction with double precision (which is slower by itself) and the DD 
changes/force imbalances, I'd say this explains the runtime issues.

-Justin

-- 
========================================

Justin A. Lemkul
Ph.D. Candidate
ICTAS Doctoral Scholar
MILES-IGERT Trainee
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================