[Pixman] [ssse3]Optimization for fetch_scanline_x8r8g8b8
xinyun.liu at intel.com
Wed Aug 11 02:00:54 PDT 2010
piman-access.c: fetch_scanline_x8r8g8b8() is mainly memcpy and 'or" operations.
With ssse3_memcpy, the performance is increased a little.
> After optimization, we can find following speed up:
> 1. fetch_scanline_x8r8g8b8()'s cycles is reduced 84% during clips run. It
> now consumes 0.8% of whole system, comparing with 2.3% w/o optimization.
> 2. system C0 is reduced around 1.5%
> 3. Global CPU utilization is reduced around 1.5% of whole system
Here is patch, please review it, thanks!
More information about the Pixman