[Pixman] [PATCH 2/3] armv7: Faster fill operations

[Matt Turner]

On Wed, Mar 4, 2015 at 5:56 PM, Ben Avison <bavison at riscosopen.org> wrote:
> This eliminates a number of branches over blocks of code that are either
> empty or can be trivially combined with a separate code block at the start
> and end of each scanline. This has a surprisingly big effect, at least on
> Cortex-A7, for src_n_8:
>
>     Before          After
>     Mean   StdDev   Mean   StdDev  Confidence  Change
> L1  1570.4 133.1    1639.6 110.7   100.0%      +4.4%
> L2  1042.6 19.9     1086.6 23.4    100.0%      +4.2%
> M   1030.8 7.2      1036.8 3.2     100.0%      +0.6%
> HT  287.4  3.5      303.3  2.9     100.0%      +5.5%
> VT  262.0  2.6      263.3  2.6     99.9%       +0.5%
> R   206.5  2.4      209.9  2.4     100.0%      +1.7%
> RT  56.5   1.0      59.2   0.5     100.0%      +4.7%
> ---

What do you use to generate this?

I'd certainly like to use it.