Ancestors

Written by Larry (Mr.Optimization) on 2025-01-31 at 10:34

Some of my work today is quite enjoyable - optimizing the open source MPEG-1 player with Arm NEON SIMD. I need to find more clients to pay me to do this.

=> More informations about this toot | More toots from fast_code_r_us@floss.social

Written by Larry (Mr.Optimization) on 2025-01-31 at 11:34

I just pushed the changes (macro block copy/add optimizations). These changes speed up video decode by another 21% overall (on my Mac M1). Now I need to rewrite them all as ESP32-S3 SIMD and see how well it performs.

=> More informations about this toot | More toots from fast_code_r_us@floss.social

Written by LovesTha🥧 on 2025-01-31 at 11:43

@fast_code_r_us 21% is pretty good for non-shit starting code.

=> More informations about this toot | More toots from LovesTha@floss.social

Toot

Written by Larry (Mr.Optimization) on 2025-01-31 at 12:04

@LovesTha The overall speedup is about 280% since I started optimizing it (pl_mpeg).

=> More informations about this toot | More toots from fast_code_r_us@floss.social

Descendants