Inference › Inference Engines › Other Engines v
tokens/sec tuning ideas

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Linear Mode
tokens/sec tuning ideas
cobalttinker Offline
Member
***
Posts: 52
Threads: 4
Joined: Sep 2026
Reputation: 0
#3
10-01-2026, 05:21 AM
What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at INT4. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.
Find
Reply


Messages In This Thread
tokens/sec tuning ideas - by latentgarden - 10-01-2026, 01:09 AM
Re: tokens/sec tuning ideas - by dockkite69 - 10-01-2026, 05:13 AM
Re: tokens/sec tuning ideas - by cobalttinker - 10-01-2026, 05:21 AM
Re: tokens/sec tuning ideas - by quiettoken - 10-02-2026, 06:00 PM
Re: tokens/sec tuning ideas - by visionforge - 10-02-2026, 07:01 PM
Re: tokens/sec tuning ideas - by fusedriver - 10-03-2026, 07:10 AM
Re: tokens/sec tuning ideas - by visionforge - 10-03-2026, 07:36 AM
Re: tokens/sec tuning ideas - by labline - 10-03-2026, 03:49 PM
Re: tokens/sec tuning ideas - by fusedriver - 10-04-2026, 04:12 PM
Re: tokens/sec tuning ideas - by llm_wren - 10-06-2026, 05:22 PM
Re: tokens/sec tuning ideas - by oxideworks59 - 10-06-2026, 04:46 AM
Re: tokens/sec tuning ideas - by cacheline - 10-06-2026, 06:25 PM
Re: tokens/sec tuning ideas - by deltastation - 10-07-2026, 06:38 AM
Re: tokens/sec tuning ideas - by mintforge - 10-10-2026, 06:56 AM
Re: tokens/sec tuning ideas - by lablab - 10-10-2026, 07:07 PM
Re: tokens/sec tuning ideas - by vectorpoint - 10-10-2026, 11:31 PM

Forum Jump:


Users browsing this thread: 1 Guest(s)