It sounds like you aren’t batching efficiently if you are being bound by memory ... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		supermatt 7 months ago \| parent \| context \| favorite \| on: Apple's MLX adding CUDA support It sounds like you aren’t batching efficiently if you are being bound by memory bandwidth.

zargon 7 months ago [–]

That’s right, in the context of Apple silicon and Halo Strix, these use cases don’t involve much batching.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact