[Sidefx-houdini-list] Performance gain in sim with different scene scale

Nikita Musatov not.a.whale at gmail.com
Tue Apr 11 06:26:53 EDT 2017


Yup. I meant it. If you change scale simulation you wanna be sure in
simulation doesn't use constant parameters, that haven't connection to
simulation with voxels, scale, particle separation or etc. Usually source
have some parameters like this.

On Tue, Apr 11, 2017, 12:42 Jordan Walsh <jordan.h.walsh at gmail.com> wrote:

> Yeah, but you can easily compensate for that as you do for changing voxel
> size per sim. You should be doing this anyway to make the sim more
> efficient.
>
> Jordan
>
> On 11 April 2017 at 10:19, Nikita Musatov <not.a.whale at gmail.com> wrote:
>
> > Simulation can be slower with small scale by source node. The source node
> > have parameters of bounding box for emitter. It can generate many unused
> > voxels.
> >
> > On Tue, Apr 11, 2017, 11:31 Olivier Jeannel <facialdeluxe at gmail.com>
> > wrote:
> >
> > > Small scale flip sims or small scale wire sim don't react the same as
> > when
> > > they are at their "regular" scale, which could lead to slower sim.
> > > If that's what you mean.
> > >
> > > On Tuesday, April 11, 2017, Jordan Walsh <jordan.h.walsh at gmail.com>
> > wrote:
> > >
> > > > So you heard from some one who heard from someone else? Hmm... I
> would
> > do
> > > > some tests if you really want to know! Turn on the performance
> monitor
> > > and
> > > > do some comparisons. You might narrow it down to a single node
> > somewhere
> > > > that is tied to scene scale. I cant say for sure, but I tend to agree
> > > with
> > > > you. Only one way to find out!
> > > >
> > > > Jordan
> > > >
> > > >
> > > >
> > > > On 11 April 2017 at 03:17, François Duchesneau <houdini at trinix.ca
> > > > <javascript:;>> wrote:
> > > >
> > > > > Hi guys,
> > > > >
> > > > > Somebody at work said he tries to avoid smaller div size simulation
> > > > > because he heard somewhere smaller scale sims tend to be slower.
> > > > >
> > > > > I don't see why it would make a difference to sim a box 1x1x1 at a
> > div
> > > > > size of 0.01 than one at 10x10x10 at 0.1.
> > > > >
> > > > > Does anyone agree with this and can somebody explain why the second
> > > case
> > > > > would be faster?
> > > > >
> > > > > Thanks
> > > > >
> > > > > Francois
> > > > > _______________________________________________
> > > > > Sidefx-houdini-list mailing list
> > > > > Sidefx-houdini-list at sidefx.com <javascript:;>
> > > > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > > > >
> > > > _______________________________________________
> > > > Sidefx-houdini-list mailing list
> > > > Sidefx-houdini-list at sidefx.com <javascript:;>
> > > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > > >
> > > _______________________________________________
> > > Sidefx-houdini-list mailing list
> > > Sidefx-houdini-list at sidefx.com
> > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> > >
> > _______________________________________________
> > Sidefx-houdini-list mailing list
> > Sidefx-houdini-list at sidefx.com
> > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
> >
> _______________________________________________
> Sidefx-houdini-list mailing list
> Sidefx-houdini-list at sidefx.com
> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>



More information about the Sidefx-houdini-list mailing list