Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
CUDA update broke my throughput with TensorRT-LLM
#4
What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at KV cache pressure. Once you know the plain path works, then add the faster path back in piece by piece.
small models can still be good
Reply


Messages In This Thread
Re: CUDA update broke my throughput with TensorRT-LLM - by mangohouse - 08-28-2026, 09:31 AM

Forum Jump:


Users browsing this thread: 1 Guest(s)