News for hackers. Stuff that matters.
Re:Apple M3 Ultra by Der_Einzige on Wednesday March 05, 2025 @01:33PM (#43270418)
No one who is using this for home use cares about anything except batch size 1 sequence size 1.
Re:Apple M3 Ultra by ein0p on Wednesday March 05, 2025 @03:30PM (#43271901)
What if you're doing bulk inference? The efficiency and throughput of bs=1 s=1 is truly abysmal.
Re:Apple M3 Ultra by saagarjha on Thursday March 06, 2025 @02:18AM (#43277368)
People want to talk to their computer, not service requests for a thousand users.