Inference › Drivers / Frameworks / Toolkits › Apple / MLX v
M3 Max for local use

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Linear Mode
M3 Max for local use
quietarch Offline
Junior Member
**
Posts: 42
Threads: 1
Joined: Oct 2026
Reputation: 0
#4
10-06-2026, 04:24 AM
Quote:Good point. The pattern around token cost has been pretty similar for me.

I'd approach it in a boring order. On Threadripper box, the easiest win is often to back off the aggressive settings and see whether the problem survives.
Find
Reply


Messages In This Thread
M3 Max for local use - by tinycontext - 10-02-2026, 03:22 PM
Re: M3 Max for local use - by tensorkite - 10-03-2026, 03:47 AM
Re: M3 Max for local use - by vectorcrate98 - 10-06-2026, 03:47 AM
Re: M3 Max for local use - by quietarch - 10-06-2026, 04:24 AM
Re: M3 Max for local use - by circuitwire58 - 10-07-2026, 05:07 PM
Re: M3 Max for local use - by torchshed - 10-08-2026, 05:09 PM
Re: M3 Max for local use - by labnest - 10-08-2026, 09:19 PM
Re: M3 Max for local use - by mlxgarden - 10-08-2026, 09:29 PM
Re: M3 Max for local use - by visionyard23 - 10-10-2026, 02:39 AM
Re: M3 Max for local use - by sanescope - 10-09-2026, 10:09 PM
Re: M3 Max for local use - by mellowstream - 10-11-2026, 03:17 PM
Re: M3 Max for local use - by amberpatch25 - 10-11-2026, 04:30 PM

Forum Jump:


Users browsing this thread: 1 Guest(s)