Re: New OS won't support 2k textures?

New Message Reply Date view Thread view Subject view Author view

Guillermo E. Gutierrez (ggutierr++at++hisd.harris.com)
Thu, 21 Jan 1999 18:23:24 -0500

Actually, we don't get any error message. The loader claims to have loaded the
texture, except when the model loads, the texture ain't there (all of the other
textures load fine)...unless, of course, Performer's ivLoader has some "hidden
callback" to turn all the texels in big textures white. :-) We haven't left
the app running more than a few minutes, so maybe it didn't get a chance to
crash...will try that tomorrow.
-^-
Guillermo E. Gutierrez "Variables are always constant
Software Engineer, Vis IR&D unless you change them."
ggutierr++at++harris.com --from the Programmer's Creed by
Harris Corporation Howard and Doran.

On Jan 22, 9:49am, Brian Corrie wrote:
> Subject: Re: New OS won't support 2k textures?
> pfHmmmm...
>
> We seem to be having a similar problem on our Onyx2 IR, 6.5.2m, Performer
> 2.2.1 eoe. Can you tell me how the Performer app is failing? We have a
terrain
> data set with a 2Kx2K texture map in an inventor file that we are trying to
> load. We get a pfSharedArena error when trying to load the file, no matter
> what we set PFTMPDIR and PFSHAREDSIZE to. Perfly just seems to keep chewing
up
> memory until it runs out and dies. Similar data sets seem to run OK if the
> texture and/or geometry is smaller.
>
> My error:
> PF Fatal/Resource: pfMemory::new() Unable to allocate 100663312
> bytes from arena 0x40000000.
> Try using pfSharedArenaSize() or env PFSHAREDSIZE
> to increase the arena size
> (currently 976562.50 KBytes) and check for adequate setrlimit()
> values and available space on swap (or pfTmpDir()).
>
> >From top (just before it dies):
> PID PGRP USERNAME PRI SIZE RES STATE TIME WCPU% CPU%
COMMAND
> 414689 414689 bcorrie 20 1059M 922M run/7 0:38 72.9 90.06
perfly_
>
> Again, strangely enough, ivview has no problem on these data sets. I suspect
> it is either an inventor loader bug or a Performer texturing bug. Resident
set
> size of ivview when it is loaded is of the order of 130MB, so the above
> numbers (RSS for perfly at 900MB+ and climbing 8-) seem way out to lunch...
>
> Anyone have any comments? Thanks in advance,
>
> Brian
>
>
>
>
>-- End of excerpt from Brian Corrie

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Thu Jan 21 1999 - 15:22:08 PST

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