Re: Non-linear transformation of pfGeoSet
Scott McMillan (mcmillan++at++cambridge.com)
Thu, 19 Feb 1998 13:01:47 -0500
On Feb 19, 2:03pm, Daniel Weiskopf wrote:
> Subject: Non-linear transformation of pfGeoSet
> Hello,
>
> I am a beginner at Performer programming. I have the following
>
> PROBLEM:
> For each frame, the vertices (coordinates and normal vectors)
> of the geometry (pfGeoSet) have to be transformed by a given
> non-linear mapping (for special relativistic visualization).
> To put it another way, this transformation can NOT be expressed
> by a normal 4x4 matrix. I'd like to achieve a frame-accurate
> behavior on a multiprocessing architecture.
>
> My ideas so far:
> The manual suggests four solutions for frame-accurate behavior:
> 1) passthrough data mechanism
> 2) frame-accurate pfSwitch
> 3) pfCycleBuffer
> 4) fluxed geosets/engines
> The transformation might be incorporated in application traversal.
>
> My question:
> Does anyone have any experience with a similar problem? In particular,
> I am interested in performance aspects/differences. Is it a good idea
> to put the transformation in the application traversal?
>
> TIA,
>
> Daniel
In a previous life, I used pfCycleBuffers/pfFlux to store arrays of vertices
and normals associated with a pfGeoSet. I then modified the vertices and
normals directly in the APP, marked the buffers as changed, and voila things
worked. I never noticed a difference in performance, but then I never bothered
to measure it either.
--
Scott McMillan
Cambridge Research Associates Voice: (703) 790-0505 x7235
1430 Spring Hill Road Fax: (703) 790-0370
McLean, VA 22102 mailto:mcmillan++at++cambridge.com
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
Submissions: info-performer++at++sgi.com
Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2.0b2
on Mon Aug 10 1998 - 17:56:48 PDT