code bloat [was Re: Semaphore assembly-code bug]

[Lee Revell]

On Mon, 2004-11-01 at 13:00 -0400, Adam Jackson wrote:
> XAA goes to a lot of effort to accelerate the primitives no one uses, and 
> desperately needs to be taken out behind the barn and shot.  KAA accelerates 
> probably a tenth of the ops XAA accelerates, but they're the right tenth so 
> the code is simpler and the performance way better.
> 
> So, yeah, chucking X is not the answer, but chucking XAA probably is.

I suspect a lot of the difference is due to better Windows drivers.  I
have an older machine with a PCI Voodoo3 card in it, one of the best
supported 3D cards under Linux, and X is blazing fast...  I wish some
new hardware was this well supported, but that is another thread.

Lee