Re: IRIX 5.3 texture repeat bug on Reality Engine

New Message Reply Date view Thread view Subject view Author view

Marcus Barnes (marcus++at++multigen.com)
Tue, 3 Oct 1995 13:54:32 -0700

On Oct 3, 2:17pm, Dorrie Hall wrote:
> Subject: IRIX 5.3 texture repeat bug on Reality Engine
>
>
> I too, had a call in about this bug a few months ago but
> the patch( I'm sorry I don't remember the number) didn't
> solve the problem on our machine. I actually had to remove
> it becuase it caused other problems. Perhaps it was only
> because I couldn't find a solution to the problem, but I've
> convinced myself the threshold that is exceeded when the
> streaking starts to appear( when I'm scaling a texture in
> Medit) is signifigant in performance terms. Things seem to
> be faster now that I've re-applied our textures so there is
> no streaking. It seems to be a nice non-abstract indicator.
> That is I think that even under 5.2 there may have been a penalty

This is true. Even though IRIX pre-5.3 handled texture repeats
correctly, it's still a RE microcode exception that degrades performance.
By modelling your textures to avoid excessive repeat counts, you
are actually staying on the RE texture "fast path" as it were ...

> for exceeding the 4 bits of repeat.
> Thanks for posting Sharon's diagnosis I never really got
> an explanation.
> Dorrie
>-- End of excerpt from Dorrie Hall

You're welcome :-)

--
Regards,
Marcus Barnes, Member Technical Staff
MultiGen Inc, 550 S. Winchester Blvd. STE 500, San Jose CA 95128
PH: 1-408-556-2654   FX: 1-408-261-4102
EMAIL: marcus++at++engr.multigen.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:51:56 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.