10-12-2026, 10:59 AM
Tried to reproduce it just now. On 5090 with LM Studio, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.
Inference
›
LLM Models
›
Open Weight
›
Recipes & Configurations
simple router policy ideas with TensorRT-LLM
simple router policy ideas with TensorRT-LLM
|
simple router policy ideas with TensorRT-LLM
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
simple router policy ideas with TensorRT-LLM - by rocmpath - 10-10-2026, 01:10 AM
Re: simple router policy ideas with TensorRT-LLM - by minthall - 10-12-2026, 01:38 AM
Re: simple router policy ideas with TensorRT-LLM - by kernelghost - 10-12-2026, 02:28 AM
Re: simple router policy ideas with TensorRT-LLM - by quantghost - 10-12-2026, 10:59 AM
Re: simple router policy ideas with TensorRT-LLM - by kernelmike - 10-15-2026, 11:23 AM
Re: simple router policy ideas with TensorRT-LLM - by rocmhouse90 - 10-15-2026, 12:21 PM
Re: simple router policy ideas with TensorRT-LLM - by quietstack - 10-15-2026, 04:48 PM
Re: simple router policy ideas with TensorRT-LLM - by bytepilot - 10-15-2026, 12:56 PM
Re: simple router policy ideas with TensorRT-LLM - by vramstream - 10-15-2026, 02:20 PM
Re: simple router policy ideas with TensorRT-LLM - by ggufwire22 - 10-15-2026, 06:06 PM
Re: simple router policy ideas with TensorRT-LLM - by labnest - 10-16-2026, 02:13 AM
|
Users browsing this thread: 1 Guest(s)