yade-users team mailing list archive
Mailing list archive
Re: R: Re: Contact detection
> After it is difficult to say if SDEC was slower or faster than Yade, I
> tried to make comparisons but it was on different computers, so it is
> not very meaningful! Please keep in mind that SDEC and Yade are made
> with different objectives. It was not very easy in SDEC to make
> modifications without change the core of the code, whereas evolutions
> and complementary packages can be easly introduced in Yade,
I didn't suggest to compare yade agains SDEC, but to compare yade
against pfc3d and esys-particle (for instance), using a well-defined
benchmark: a funnel with 10k,100k,1M particles for something that moves
and perhaps isotropic compression for a more-or-less static simulation.
> everybody can take advantage of developments made by others.
Oh, really. I keep rewriting bad code all the time :-|
> my conclusion: speed is not the only key point.
I don't buy this, if that means: we don't have to care about speed. For
research, flexibility is good, but if it is at the cost of significantly
lower speed when doing simulations, than something's wrong.