From: Christian Rothländer (Mailinglisten++at++rothlaender.net)
Date: 01/18/2006 05:37:40
Hi there
Gordon Tomlinson wrote:
> Intersection test are expensive especially if you do not have a well
> balanced tree (quad or similar) and or large pools geometry and you can only
> do a limited number per frame depending on you scenes complexity,
> composition and desired frame rate etc
The Graph is made by a biologist, it don't think it's well balanced or even something balanced if that don't happen by
pure chance...
Another idea was to shoot the intersection lines from the "Objects of Interest" to the viewers position, but they didn't
liked it because the didn't want to change something on the scene graphs .. So I have to work "brute force" and search :-(
> In the first instance your intersections code should be testing bounding
> scene of nodes so that it can quickly reject parts of the scene graph that
> are not with in you line segments, that why a well balanced and spatially
> partition tree is so important for intersection test. If you data is big
> soups of tri's then you need to look at breaking them up into discreet
> spatially partitioned nodes etc
>
> The GPU will know nothing about your scene graph so you will get back very
> limited hit information how would you know which node had been hit etc,
> intersections testing is inherently tied to your scene graph and the
> majority of case.
I realized that after I wrote my mail :-(
>
> One way to typically way to gain extra performance with isectors is to
> run-them asynchronously and locked on to a separate CPU
I don't think they want to buy new hardware for that problem, the analysis is working offline, I just have to reach the
goal: "we start it on the evening, and the next morning we'll have the results"
> Another approach is to run the intersections on a dedicate host machine that
> simple has a copy of the scene graph for intersection and does not have nay
> visuals or draw etc
I tried to deactive the graphical system, but it didn't went much faster.
> 90x ? What?
It's for an eye-testing experiment, the person you are testing sits with a joystick about 30min in front of the canvas.
With the actual benchmark the processing time afterwards is 90 * 30 min = nearly 2 Days
>
> 600000 tests with what? is this triangles/primitives or 600000 isectors
I have to call that function 600.000 times per second
I hope this information makes the problem better to understand
Christian
This archive was generated by hypermail 2b29 : Wed Jan 18 2006 - 05:38:10 PST