yade-dev team mailing list archive

[Janek Kozicki <janek_listy@xxxxx>]

Bruno Chareyre said:     (by the date of Fri, 22 Feb 2008 14:24:12 +0100)

> I think speed should be kept in mind with high priority when 
> implementing new serialization and new container.

I expect some speed improvement here. If it will be slower that will
be an unpleasant surprise.


> Hehe, I think you can't develop in a really approriate way if you are 
> not a user at the same time.

true :-)


> Speed, for instance, is one of the major issue for most users.

I still prefer to look at design quality, because with good quality
we can win much more speed than with only "optimizations".

For example going parallel in a "more optimized" design like
in /extra/spherical-dem-simulator/ is much more difficult.
So SphericalDEMSimulator.cpp is currently faster because it was
totally optimized, but it cannot use the power of multiprocessor
computations. This is just one example, but I hope that you grasp
what I mean here.

Of course those decisions must be balanced, so that's why we are
having this discussion. It is good. And of course it's good that I'm
yade user too.

> >> I could run the triaxial test around 25% faster.

congrats!

> >> I can commit those changes soon, I'm curious to know the results on 
> >> other computers.

Go on and commit. I suppose that you already checked on my  computer.


best regards
-- 
Janek Kozicki                                                         |
_______________________________________________
yade-dev mailing list
yade-dev@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-dev