Alexander Schmidt wrote:
> Hi Mesa team,
> 
> I have read on hpc-consortium.net that there is a work-in-progress
> project on porting Mesa to the Cell processor.

That information isn't really accurate.  I attended the Cell 
"Hack-a-thon" back in February.  It was a week-long workshop to learn 
Cell programming.  Michael McCool and I spent a few days incorporating 
his RapidMind toolkit into Mesa's fragment program interpreter.

At the end of the week we had enough working that the 'fslight' demo ran 
about 7 times faster (than just running on the PPC core) but it was 
still pretty slow.  Since a limited number of instructions were 
supported, that was the only demo that actually worked.  We haven't done 
any more work since then.


> As I am currently looking
> for a diploma thesis topic dealing with the cell and porting software
> to it, I would like to know how the status of this project is and if
> there are work items left where you could need some support.

The RapidMind toolkit isn't the right solution for a real OpenGL 
implementation on Cell, for various reasons.  To get decent performance, 
more radical architectural changes will be needed.  We (TG) are hoping 
to do some work in this area in the future.

Is your thesis more concerned with "just porting" to Cell, or writing 
highly-tuned Cell software?  My impression is that getting the best 
performance out the Cell processor often involves re-architecting 
software from the bottom up.

-Brian

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Mesa3d-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mesa3d-dev

Reply via email to