09-28-2026, 11:16 PM
Can you fill in a bit more context? What model, what runtime, and roughly what context size?
|
Blackwell deployment chatter
|
|
09-28-2026, 11:16 PM
Can you fill in a bit more context? What model, what runtime, and roughly what context size?
09-29-2026, 07:48 AM
Quote:Can you fill in a bit more context? What model, what runtime, and roughly what context size? 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.
Mostly CUDA, sometimes wrong, usually measured
09-29-2026, 08:31 AM
that fixed it for me too.
09-29-2026, 09:11 AM
I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.
09-29-2026, 11:44 AM
Tried to reproduce it just now. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.
If you post your exact settings, people here can probably narrow it down pretty fast. |
|
« Next Oldest | Next Newest »
|