imlib2/xrender benchmarks

[Owen Taylor]

Cameron wrote:
> Hello,
> 
> I'm sure many people have seen imlib2 vs xrender benchmarks before,
> however, I still think these deserve a look. I do have render
> acceleration enabled for my card. My card is also quite high end.
> 
> 
> http://audioseek.net/~hiryu/render.html
> 
> Anyway, hope this is helpful.

Note that in by far the most common operation (non-scaled OVER)
RENDER came out ahead by a large margin. For the scaled blends
note that that:

  - Virtually all drivers don't yet accelerate scaled RENDER blends.
    (I think that's *all* drivers from the xorg tree)

  - There's no fast-pathing of scaled RENDER blends in software at
    all. It's all slow path.

So the results are absolutely not suprising. Nobody is going to
object if you help fix that situation.

Really, fast routines to implement one operation or another are
a very well defined problem. The hard thing is getting the framework
in place so that we have the right client/server interface (RENDER),
and so that the software code lives in only one place, whether
it's used by the client or the server. (libpixman)

I think that's coming together now.

Regards,
						Owen