yade-users team mailing list archive

[Jan Stránský <question219884@xxxxxxxxxxxxxxxxxxxxx>]

Question #219884 on Yade changed:
https://answers.launchpad.net/yade/+question/219884

Jan Stránský proposed the following answer:
Hello Jessica,

it may depend on particle radius in certain sense..
you can estimate the time of the simulation like this:

totalTime = numberOfSteps * computationTimeOfOneStep

numberOfSteps can be determined (in your case) from strain rate and
timeStep dt ( numberOfSteps = finalStrain / (strainRate * dt) )

dt is dependent on mass, stiffness and size of your particles (see e.g.
[1]).

computationTimeOfOneStep is dependent on number of particles (approximately
N*log(N) ) and power of your computer. Usually it is not difficult to
measure it (do not measure the very first step as it is O(N^2) due to the
collider's  first sorting of particle positions). So if you know
computation time for certain number of particles, you can estimate
computation time for different number of particles. You can try to use
these formulas to estimate the time of your simulation and it should not
*dramatically* differ from reality.

hope I wrote it in understandable way :-)
cheers
Jan

[1] https://yade-
dem.org/doc/yade.utils.html#yade.utils.SpherePWaveTimeStep

-- 
You received this question notification because you are a member of
yade-users, which is an answer contact for Yade.