10-18-2026, 07:45 PM
Tried to reproduce it just now. On GB200 with LM Studio, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.
VRAM first, aesthetics later
|
4090 vs 5090 for 70B quants with TensorRT-LLM
|
|
« Next Oldest | Next Newest »
|
Users browsing this thread: 1 Guest(s)