09-25-2026, 05:29 PM
I tested this tonight. On B200 with TensorRT-LLM, I saw basically the same behavior. The main thing that moved was latency, not quality.
still benchmarking everything
Inference
›
Training & Customization
›
Model Merging
vector DB that stayed out of the way on dual 3090
vector DB that stayed out of the way on dual 3090
|
vector DB that stayed out of the way on dual 3090
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
vector DB that stayed out of the way on dual 3090 - by nanogarden - 09-21-2026, 09:35 PM
Re: vector DB that stayed out of the way on dual 3090 - by benchplease - 09-22-2026, 10:21 AM
Re: vector DB that stayed out of the way on dual 3090 - by paperrunner - 09-22-2026, 03:02 PM
Re: vector DB that stayed out of the way on dual 3090 - by vramstream - 09-25-2026, 03:44 PM
Re: vector DB that stayed out of the way on dual 3090 - by quietcraft - 09-25-2026, 05:29 PM
Re: vector DB that stayed out of the way on dual 3090 - by sparrowshed - 09-25-2026, 06:13 PM
Re: vector DB that stayed out of the way on dual 3090 - by phasespark - 09-25-2026, 06:20 PM
Re: vector DB that stayed out of the way on dual 3090 - by tinycontext - 09-26-2026, 06:51 AM
|
Users browsing this thread: 1 Guest(s)