• 0 Posts
  • 28 Comments
Joined 1 year ago
cake
Cake day: June 23rd, 2023

help-circle
  • Google is pushing av1 because of patents, but 266 is just plain better tech, even if it’s harder to encode.

    This same shit happened with 265 and vp9, and before that, and before that with vorbis/opus/aac.

    They’ll come back because it’s a standard, and has higher quality.

    Maybe this is the one time somehow av1 wins out on patents, but I’m encoding av1 and I’m really not impressed, it’s literally just dressed up hevc, maybe a 10% improvement max.

    I’ve seen vvc and it’s really flexible, it shifts gears on a dime between high motion and deep detail, which is basically what your brain sees most, while av1 is actually kind of worse than hevc at that to me, it’s sluggish at the shifts, even if it is better overall.












  • So, trying not to dox myself, I worked with the architect twice.

    Knights Ferry was derived directly from Larrabee (GPU), P54Cs with pre-AVX-512, .

    KNC was a die shrink with more cores. Both of these were PCIe accelerators only.

    KNL had full Airmont Atom cores with smt4, basically meaningful cores with proper AVX-512. Also you could boot them with linux, or as a PCIe accelerator.

    KNM jadded ML instructions, basically 8/16bit float and faster SIMD.

    They cancelled KNH.

    I interviewed some of the actual Larrabee guys, they were wild, there was a lot of talk about dynamic translation, they were trying to do really complex things, but when people talk like that it makes me think they were floundering on the software and just looking for tech magic solutions to fundamental problems.

    Intel always dies when the software gets more complex than really simple drivers, it’s their achilles heel.

    KNL also had the whole MCDRAM on package for basically HBM bandwidth, but that didn’t actually work very well in practice, again due to software issues (you have to pick where you allocate, and using it as an l4 cache was not always effective).