arm: vp9mc: Minor adjustments from review of the aarch64 version
authorMartin Storsjö <martin@martin.st>
Thu, 10 Nov 2016 09:07:39 +0000 (11:07 +0200)
committerMartin Storsjö <martin@martin.st>
Thu, 10 Nov 2016 09:18:22 +0000 (11:18 +0200)
commit557c1675cf0e803b2fee43b4c8b58433842c84d0
tree4cb2063cbbc569fc45faf05ae35d3c66fe29e69b
parent383d96aa2229f644d9bd77b821ed3a309da5e9fc
arm: vp9mc: Minor adjustments from review of the aarch64 version

This work is sponsored by, and copyright, Google.

The speedup for the large horizontal filters is surprisingly
big on A7 and A53, while there's a minor slowdown (almost within
measurement noise) on A8 and A9.

                            Cortex    A7        A8        A9       A53
orig:
vp9_put_8tap_smooth_64h_neon:    20270.0   14447.3   19723.9   10910.9
new:
vp9_put_8tap_smooth_64h_neon:    20165.8   14466.5   19730.2   10668.8

Signed-off-by: Martin Storsjö <martin@martin.st>
libavcodec/arm/vp9dsp_init_arm.c
libavcodec/arm/vp9mc_neon.S