Re: MPU

New Message Reply Date view Thread view Subject view Author view

Brian Cabral (cabral++at++sgi.com)
Wed, 26 Aug 1998 22:04:32 -0700

> Date: Wed, 26 Aug 1998 18:49:42 -0700
> From: george williams <gcw++at++best.com>
>
>
> I hope questions about MPU (MultiPipe Utility) are not out-of-line for
> this newsgroup. But, having just read the MPU spec, I am wondering if
> any Performers out there have used it. It would *seem* that MPU is
> equal to Performer minus the Performer scene graph. Is this a fair
> observation?
>

Not quite :)

MPU, forks your app replicating in different address spaces your scene
graph or any other retained graphics data structure. This means it
does not share the scene graph in a multi-buffered arena. This in turn
requires means the two scene graphs must be synched on a per frame
basis -taking up twice the memory footprint. Additionally the two
processes' windows do not de-mux input events into a single event
queue as Performer does.

MPU provides a great mechanism to help openGL apps that are not
multi-pipe become so. But it only addresses immediate mode rendering
multiplexing.

-------------------------------------------------------------------------
Brian Cabral cabral++at++sgi.com
Director - Advanced Graphics Software 8U-590
Silicon Graphics Computer Systems tel: 650-933-2694

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Wed Aug 26 1998 - 22:04:37 PDT

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