Slacker News Slacker News logo featuring a lazy sloth with a folded newspaper hat
  • top
  • new
  • show
  • ask
  • jobs
Library
← Back to context

Comment by Readerium

8 hours ago

LLMs are memory bandwidth bound not compute bound.

3 comments

Readerium

Reply

ondra  7 hours ago

This is incorrect, prompt processing is compute bound.

AntiUSAbah  5 hours ago

LLMs are bound by both and depends on the hardware which factor is higher.

icelancer  6 hours ago

This is only true for some parts of the time cost function.

Slacker News

Product

  • API Reference
  • Hacker News RSS
  • Source on GitHub

Community

  • Support Ukraine
  • Equal Justice Initiative
  • GiveWell Charities