[FFmpeg-devel] [PATCH 11/11] avcodec/lossless_videodsp: add AVX-512 version of add_bytes

James Darnley jdarnley at obe.tv
Fri Nov 10 15:30:45 EET 2017


On 2017-11-09 20:43, Martin Vignali wrote:
> 2017-11-09 20:37 GMT+01:00 Martin Vignali <martin.vignali at gmail.com>:
>> lgtm
>>
>> Can you post your checkasm benchmark result for this ?

Yep

> $ ./tests/checkasm/checkasm --bench --test=llviddsp
> benchmarking with native FFmpeg timers
> nop: 26.0
> checkasm: using random seed 3684557040
> SSE2:
>  - llviddsp.add_bytes             [OK]
>  - llviddsp.add_median_pred       [OK]
> SSSE3:
>  - llviddsp.add_left_pred_zero    [OK]
>  - llviddsp.add_left_pred_rnd_acc [OK]
>  - llviddsp.add_left_pred_int16   [OK]
> SSE4.1:
>  - llviddsp.add_left_pred_int16   [OK]
> AVX2:
>  - llviddsp.add_bytes             [OK]
> AVX-512:
>  - llviddsp.add_bytes             [OK]
> checkasm: all 8 tests passed
> add_bytes_c: 701.0
> add_bytes_sse2: 19.0
> add_bytes_avx2: 78.0
> add_bytes_avx512: 10.0
> add_left_pred_int16_c: 3324.5
> add_left_pred_int16_ssse3: 2360.5
> add_left_pred_int16_sse4: 797.5
> add_left_pred_rnd_acc_c: 2074.0
> add_left_pred_rnd_acc_ssse3: 461.5
> add_left_pred_zero_c: 1987.0
> add_left_pred_zero_ssse3: 461.5
> add_median_pred_c: 15809.5
> add_median_pred_sse2: 1113.5



More information about the ffmpeg-devel mailing list