yade-dev team mailing list archive

[Jerome Duriez <Jerome.Duriez@xxxxxxxxxxx>]

Hi,

General discussions apart, what is the problem with this commit ?


________________________________
From: Yade-dev [yade-dev-bounces+jerome.duriez=ucalgary.ca@xxxxxxxxxxxxxxxxxxx] on behalf of Bruno Chareyre [bruno.chareyre@xxxxxxxxxxxxxxx]
Sent: December 3, 2014 1:27 PM
To: yade-dev@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Yade-dev] [Branch ~yade-pkg/yade/git-trunk] Rev 3486: Parallelization of interaction loops in Law2_..._Capillarity (which is in fact not a LawFunctor h...

Hi Jérôme,
I was reviewing commits of last weeks and I see this thing.
1/ Very important: please don't try things like that without getting approval here on yade-dev from at least the author of the code (and not by a private message to the author, if this is what you did) and/or other devs. You are not expert in openMP, right? Parallelization is very dangerous, didn't you feel like you could maybe introduce horrible bugs?! The fact that it works for one or two scripts doesn't mean that it is bug free.

2/ Don't waste too much of your time, as mentionned before this code will disappear.

Cheers

Bruno


On 17/10/14 22:00, noreply@xxxxxxxxxxxxx<mailto:noreply@xxxxxxxxxxxxx> wrote:

------------------------------------------------------------
revno: 3486
committer: jduriez <jerome.duriez@xxxxxxxxxxx><mailto:jerome.duriez@xxxxxxxxxxx>
timestamp: Fri 2014-10-17 09:43:56 -0600
message:
  Parallelization of interaction loops in Law2_..._Capillarity (which is in fact not a LawFunctor handled by InteractionLoop). Moreover adopting the FOREACH iterator for the non-parallel flavour.

  Considering simulations of a dense sample with 10000 bodies, executions of capillary Law2 took :
  - before change : ~ 1700 s / 100 000 executions, whatever j option (j4 or j1 tested)
  - after change : 620 s / 100 000 executions with j4 (only one run tested)

  So:
  - Law2_ScGeom_CapillaryPhys_Capillarity is now parallelized (thanks to InteractionLoop example...)
  - approx *3 speedup is reached from j1 to j4 (better benchmark could maybe have been done)
modified:
  pkg/dem/Law2_ScGeom_CapillaryPhys_Capillarity.cpp


--
lp:yade
https://code.launchpad.net/~yade-pkg/yade/git-trunk

Your team Yade developers is subscribed to branch lp:yade.
To unsubscribe from this branch go to https://code.launchpad.net/~yade-pkg/yade/git-trunk/+edit-subscription




_______________________________________________
Mailing list: https://launchpad.net/~yade-dev
Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx<mailto:yade-dev@xxxxxxxxxxxxxxxxxxx>
Unsubscribe : https://launchpad.net/~yade-dev
More help   : https://help.launchpad.net/ListHelp




--
_______________
Bruno Chareyre
Associate Professor
ENSE³ - Grenoble INP
Lab. 3SR
BP 53
38041 Grenoble cedex 9
Tél : +33 4 56 52 86 21
Fax : +33 4 76 82 70 43
________________