2 comments

[ 2.7 ms ] story [ 72.2 ms ] thread
Full tweet:

> SpaceX has almost finished writing V1.0 of an in-house AI training stack in C that exact-maps to 220k GB300s with 800G NICs, making heavy use of pipeline parallelism and getting as close to bare metal as possible.

> The potential speed improvement vs JAX for large training runs is over an order of magnitude