|
|
John VanSickle wrote:
>
> Matt Giwer wrote:
> >
> > For some time I have been considering how to make extremely complex
> > shapes such as Babylon 5. I considered on huge mesh but that is hugely
> > memory intensive and has smoothing problems.
>
> Smooth triangles take up something like 144 bytes each when they are
> part of a mesh.
>
> The Greb (one of my regular IRTC models) is made of ten different
> meshes, some used once, some used twice, and a couple used four times.
> Together there are over thirty thousand triangles in the Greb, but due
> to mesh copying only 13,376 need be kept. Together the meshes take up
> around 2 Megs of ram, and if I add Greb, it takes only a little bit
> mopre RAM.
Yes, but ...
On binaries I have posted several examples of what I am talking about
with the images and migrating from meshes to heighfields for problems of
computational loading with an acceptable 333/128 PII linux machine. I
produced 13 MB mesh file to render. The console had no problem but KDE
graphics window did. I do not explain that, I only observe it.
That said, I don't know what the hell is going on. I documented it all
and with source code and that is what happens.
I was running about 256x256x2 triangles and X11/KDE on linux was
unhappy but a console terminal ran it quite well than thank you. It is
really something I have to document and put before both the X11 and KDE
folks and let them fight it out, not me. Linux (Redhat) per se is
totally stable without their add-ons.
--
In the 19th century the British decided they were
naturally fit to rule the world. So much for the
rulers of the world.
-- The Iron Webmaster, 143
Post a reply to this message
|
|