I'm creating the base mesh outside the Iterator, then passing it in, and adding per-vertex colour with the same OpenCL kernel that distorts the mesh, so I guess it's a different setup.
a|x
--- On Wed, 4/11/09, George Toledo <email@hidden> wrote:
> From: George Toledo <email@hidden>
> Subject: Re: iterator performance
> To: "Alex Drinkwater" <email@hidden>
> Cc: "Christopher Wright" <email@hidden>, "quartzcomposer-dev list list" <email@hidden>
> Date: Wednesday, 4 November, 2009, 21:21
> Not to sidetrack, but for clarity on that
> issue, placing the mesh render patch in an iterator breaks
> the color values created by the mesh creator outside of the
> iterator. To be clear, if the mesh creator isn't
> actually setup to create color value - no structure input on
> the "Colors" input - everything works
> "ok", and one can do all the typical iterator
> color stuff by using the color control on the actual mesh
> render. I'm venturing a guess that you aren't
> creating your color values with javascript input into the
> "Colors" input of the mesh creator and then
> iterating the mesh render (when I "unhook" that, I
> can at least control color on the mesh render inside of the
> iterator). Let me know if you are (offlist or not), because
> it may lead to a valid workaround for a bug that I need to
> file.
>
> -George Toledo
> On Wed, Nov 4, 2009 at 4:02 PM,
> Alex Drinkwater <email@hidden>
> wrote:
>
> Funnily
> enough, I have managed to get generative mesh stuff working
> with an Iterator and OpenCL, with different coloured mesh
> per-Mesh Renderer instance. It's not as fast as I would
> like, but since it uses OpenCL, there's not way to
> compare with QC 3.
>
>
>
>
> a|x
>
>
>
>
>
> --- On Wed, 4/11/09, George Toledo <email@hidden>
> wrote:
>
>
>
> > From: George Toledo <email@hidden>
>
> > Subject: Re: iterator performance
>
> > To: "Christopher Wright" <email@hidden>
>
> > Cc: "quartzcomposer-dev list list" <email@hidden>
>
> > Date: Wednesday, 4 November, 2009, 20:12
>
> > I'm getting the
> same basic fps results
>
> > using the same hardware to test both qtz's -
> performance
>
> > is pretty much halved in SL (as in, half as fast). I
> could
>
> > post more examples, but you've covered it.
>
> >
>
> > I haven't found any scenarios where the new
>
> > iterator is quicker, and the iterator also always
> seems to
>
> > be a problem patch in scenarios where other things
> break,
>
> > though it may just be coincidence.
>
> >
>
> > Off the top of my head, iterator+shadows are a no-go,
>
> > iterator + mesh creator results in the color values
> from the
>
> > mesh creator getting broken (still need to file bug
>
> > officially), and a particular queue+iterator thing
> that
>
> > worked for me in 10.5 no longer works. This could all
> be
>
> > side issues to this performance problem, but when
> taken
>
> > together, it makes me dubious about the new iterator,
> in
>
> > general.
>
> >
>
> > -George Toledo
>
> >
>
> > On Wed, Nov 4, 2009 at 2:48 PM,
>
> > Christopher Wright <email@hidden>
>
> > wrote:
>
> >
>
> > <puts on asbestos suit ;)>
>
> >
>
> > I'm finding a lot of cases where iterator
> performance
>
> > in QC4 is dramatically worse than in QC3 (typically
> about
>
> > half as fast).
>
> >
>
> >
>
> >
>
> >
>
> > (this does a 50x50 grid, with half the sprites
> disabled,
>
> > so it's 2500/2 = 1250 sprites per frame -- MacBook
> can
>
> > render this at 60fps in Leopard, MacBook Pro is
> struggling
>
> > to hit 30fps in Snow Leopard.)
>
> >
>
> >
>
> >
>
> >
>
> > (this one's iteration count is cranked up
> really
>
> > high, but it's still ~5fps on my MacBook
> 1.83GHzGMA950
>
> > running 10.5.8, and only 2.2-2..5fps on my MacBook Pro
> 2.4GHz
>
> > 8600M GT running 10..6.x (the MacBook on Snow Leopard
> also
>
> > hits about 2.5fps, so the same machine is half as fast
> in
>
> > iterators). Turning down the iteration count makes
> it
>
> > scale up, with 10.5 always running about 2x as fast
> as
>
> > 10.6).
>
> >
>
> >
>
> >
>
> >
>
> > Is this expected? Does anyone else experience this
> kind
>
> > of performance? Are there cases where iterators are
>
> > actually faster (or at least as fast) on Snow
> Leopard?
>
> > I'm only seeing a few cases where having all
> consumer
>
> > subpatches get disabled causes snow leopard's
> iterator
>
> > to skip work (unlike leopard's). Should I file a
> bug
>
> > for this?
>
> >
>
> >
>
> >
>
> >
>
> > --
>
> >
>
> > [ christopher wright ]
>
> >
>
> > email@hidden
>
> >
>
> > http://kineme.net/
>
> >
>
> >
>
> >
>
> >
>
> > _______________________________________________
>
> >
>
> > Do not post admin requests to the list. They will be
>
> > ignored.
>
> >
>
> > Quartzcomposer-dev mailing list (email@hidden)
>
> >
>
> > Help/Unsubscribe/Update your Subscription:
>
> >
>
> > http://lists.apple.com/mailman/options/quartzcomposer-dev/email@hidden
>
> >
>
> >
>
> >
>
> > This email sent to email@hidden
>
> >
>
> >
>
> >
>
> > -----Inline Attachment Follows-----
>
> >
>
> >
>
> > _______________________________________________
>
> > Do not post admin requests to the list. They will be
>
> > ignored.
>
> > Quartzcomposer-dev mailing list (email@hidden)
>
> > Help/Unsubscribe/Update your Subscription:
>
> > http://lists.apple.com/mailman/options/quartzcomposer-dev/email@hidden
>
>
> >
>
> > This email sent to email@hidden
>
>
>
>
>
>
>
>
>
>
>
>
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Quartzcomposer-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/quartzcomposer-dev/email@hidden
This email sent to email@hidden
