FFmpeg devs boast of another 100x leap thanks to handwritten assembly code

Re:FFmpeg devs boast of another 100x leap thanks to handwrit... by brigade on Sunday July 20, 2025 @09:12PM (#44630907)

It's AVX512 that makes the gains, not assembly. This kernel is simple enough that it wouldn't be measurably faster than C with AVX512 intrinsics.

And it's 100x because a) min/max have single instructions in SIMD vs cmp+cmov in scalar and b) it's operating in u8 precision so each AVX512 instruction does 64x min/max. So unlike the unoptimized scalar that has a throughput under 1 byte per cycle, the AVX512 version can saturate L1 and L2 bandwidth. (128B and 64B per cycle on Zen 5.)

But, this kernel is operating on an entire frame; if you have to go to L3 because it's more than a megapixel then the gain should halve (depending on CPU, but assuming Zen 5), and the gain decreases even more if the frame isn't resident in L3.

[ Parent | on HN ]

Re:FFmpeg devs boast of another 100x leap thanks to handwrit... by saati on Sunday July 20, 2025 @10:11PM (#44631166)

The AVX2 version was still 64x faster than the C one, so AVX-512 is just 50% improvement over that. Hand vectorized assembly is very much the key to the gains.

[ Parent | on HN ]

Re:FFmpeg devs boast of another 100x leap thanks to handwrit... by brigade on Sunday July 20, 2025 @10:53PM (#44631357)

The only material difference AVX2 makes is that it can't saturate L1 bandwidth. Which would imply that 100x for AVX-512 is only for frames that fit within L1.

And... yep, the benchmark on 256x16 frames. [1]

[1] https://ffmpeg.org/pipermail/ffmpeg-devel/2025-July/346729.h...

[ Parent | on HN ]