[gmx-users] Load imbalance

[Justin A. Lemkul]

Carla Jamous wrote:
> Yes, but I forgot to say that I got this note that followed:
> 5.4 % performance was lost because the PME nodes
>       had less work to do than the PP nodes.
>       You might want to decrease the number of PME nodes
>       or decrease the cut-off and the grid spacing.
> 

5% loss still isn't that bad.  I've achieved good results with worse loss :)  It 
all depends (again) on your setup.  When you built the .tpr file, grompp should 
have reported what it predicted for the % of the load that PME should be 
calculating.  For most cases, it will be best if it is 25-33%, such that the 
PP/PME ratio can be 3:1 or 2:1.  If you don't achieve absolute perfection (which 
is rare), you're going to experience some small performance loss.  You can (as 
the message is indicating to you) tweak some of the input parameters, such as 
the Fourier grid spacing, to try to re-balance things.

Again, much of this is in the manual and the Gromacs 4 paper.

-Justin

> Carla
> 
> On Fri, Jan 22, 2010 at 1:34 PM, Erik Marklund <erikm at xray.bmc.uu.se 
> <mailto:erikm at xray.bmc.uu.se>> wrote:
> 
>     Carla Jamous skrev:
> 
>         Hi everyone,
> 
>         Lately I've been using gromacs version 3.3.3. Yesterday, I
>         started simulations in a cluster where gromacs version is 4.0.3.
>         After a run, I got this error message:
>          Average load imbalance: 0.5 %
>          Part of the total run time spent waiting due to load imbalance:
>         0.3 %
>          Steps where the load balancing was limited by -rdd, -rcon
>         and/or -dds: X 0 % Y 0 %
>          Average PME mesh/force load: 0.898
>          Part of the total run time spent waiting due to PP/PME
>         imbalance: 2.7 %
> 
>         I don't have a clue what this means & how can I fix it?
> 
>         Please can anyone help?
> 
>         Thank you
>         Carla
> 
>     This is no error message. 0.5 % load imbalance just means that the
>     workload was distributed in a good way over procesors throughout
>     your simulation.
> 
>     -- 
>     -----------------------------------------------
>     Erik Marklund, PhD student
>     Laboratory of Molecular Biophysics,
>     Dept. of Cell and Molecular Biology, Uppsala University.
>     Husargatan 3, Box 596,    75124 Uppsala, Sweden
>     phone:    +46 18 471 4537        fax: +46 18 511 755
>     erikm at xray.bmc.uu.se <mailto:erikm at xray.bmc.uu.se>  
>      http://xray.bmc.uu.se/molbiophys
> 
>     -- 
>     gmx-users mailing list    gmx-users at gromacs.org
>     <mailto:gmx-users at gromacs.org>
>     http://lists.gromacs.org/mailman/listinfo/gmx-users
>     Please search the archive at http://www.gromacs.org/search before
>     posting!
>     Please don't post (un)subscribe requests to the list. Use the www
>     interface or send it to gmx-users-request at gromacs.org
>     <mailto:gmx-users-request at gromacs.org>.
>     Can't post? Read http://www.gromacs.org/mailing_lists/users.php
> 
> 

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

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

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