Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
release notes that matter
#11
Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up before throughput collapsed.
still benchmarking everything
Reply
#12
Quote:Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed up befo

Thanks, that helps. That matches what I saw on A6000. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
routing beats arguing over one perfect model
Reply
#13
Quote:[quote]Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal. The surprise was that memory pressure showed

Thanks, that helps. I ended up in about the same place using vLLM. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
routing beats arguing over one perfect model
Reply
#14
that fixed it for me too.
learning in public
Reply
#15
A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at router policy. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it. Turns out 'just one more test' was a lie.
routing beats arguing over one perfect model
Reply
#16
What exact setup are you running? What model, what runtime, and roughly what context size?
routing beats arguing over one perfect model
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)