Just a note that I’ve committed the bulk of a functional tcl radeon driver tothe new ‘tcl-0-0-branch’. This is basically my code as of about a 10 daysago, which exercises most of the t&l hardware but doesn’t have any cooloptimizations. Since then I’ve been working on cool optimizations, butthey’re not ready yet.
I’m now running a compile on what I committed & there will probably be a fewmore commits before it’s all up & running. I’ll post when I think it’s in areasonable state.
Then a few hours later:
OK. Things seem to be more or less working now. Over the next week or soI’ll be finishing and checking in code for the following:
- Array (rather than vertex) based dma of vertex data. This will probablyhelp q3 performance.
– Fast code for immediate (begin/vertex/end) mode rendering.
Until then, people who are interested should feel free to check the code outand have a play. I don’t expect it to be too robust yet, but bug reports arewelcome.
Anybody doing development on the Radeon (hi Michael, Ian, others) shouldprobably work off this codebase as it is somewhat divergent from what’s on thetrunk in terms of the way state is managed. I’d very much hope to get thisstabilized and merged sooner rather than later.
A big thanks to ATI for allowing us to release this code and to TungstenGraphics for having the vision to persue it.
Thanks to Joe Drew for the heads up.