SIMD
Single Instruction Multiple Data; a class of CPU instructions that perform one operation on a whole vector of values at once, well suited to pixel grids in video.
As defined in Lex Fridman Podcast #496 - FFmpeg
Where it comes up · 1
Most important claims · 3 of 3Search all claims for “SIMD” →
- FFmpeg/dav1d achieve extreme speed by abusing the machine in unintended ways, including using cryptography instructions for unrelated video processing and bypassing the operating system's function calling convention to create custom calling conventions within their own libraries, avoiding costly register save/restore through L1/L2 cache.Lex Fridman Podcast #496 - FFmpeg
- Handwritten SIMD (single instruction multiple data) assembly delivers 10x-50x (sometimes 62x) speedups over scalar code and is multiple times faster than compiler auto-vectorization or intrinsics, contradicting the common belief that good compilers can match handcrafted assembly; this matters because FFmpeg/dav1d run on order-of-magnitude a billion CPUs where every cycle counts.Lex Fridman Podcast #496 - FFmpeg
- Working on FFmpeg/VLC is the best programming school because it demands real-time correctness (16ms per frame, no margin), requires understanding computer architecture (CPU pipelining, SIMD, ALU, I/O), and provides code reviews from the most seasoned programmers, making contributors among the best programmers even if they later write TypeScript.Lex Fridman Podcast #496 - FFmpeg