Re: Fwd: Re: pfLightpoint alpha
John Rohlf (jrohlf++at++tubes)
Fri, 10 Mar 95 17:41:19 PST
>
> I don't think that 1000 extra (invisible) light points are worth the expense
> of checking every light point to see if its alpha value is close to 0. (Via
> some PF_ALMOST_EQUAL function with a certain tolerance threshold). It takes
> about 4-5ms on an Indigo2 Extreme to display 5000 light points (all same size,
> but varying colors & alpha values). I was thinking about making each light
> point (a star in my case) be a different size, but I haven't gotten around to
> it yet. It would require a different light point node for every size, or for
> every point, which might be a serious drain on memory and cause performance
> loss. Any comment on how (or if it is worth it) to make varying or random
> sizes for about 5000 light points?
> ______________________________________________________________________________
>
> Just to add another question to the discussion:
>
> I've been using plain GL to create a background starfield effect - is this more,
> or less efficient than using light points if they're all the same size & color?
>
> Would GL be better in Lance's case (above) for varying sizes?
I suggest you use a pfGeoSet instead of pfLightPoint for stars.
I think a pfGeoSet would provide adequate performance.
This archive was generated by hypermail 2.0b2
on Mon Aug 10 1998 - 17:51:03 PDT