[x264-devel] [PATCH 3/3] arm64: intra_predict_4x4_h: use multiplication with 0x01010101 for splat

[Henrik Gramner]

2015-08-18 10:25 GMT+02:00 Janne Grunau <janne-x264 at jannau.net>:
> On a cortex-a53:
>                      gcc 4.9.2   llvm 3.6   neon (before)   neon (after)
> intra_predict_4x4_h: 162         147        160/155         139/135
> ---
>  common/aarch64/predict-a.S | 29 +++++++++++++----------------
>  1 file changed, 13 insertions(+), 16 deletions(-)

Pushed all 3 to sandbox. Thanks.