2026-11-01 · America/Los_Angeles · 社区动态 · #3

Getting incredible results with a single RTX 3090ti and 128gb of ram. Deepseek-V4.1-Flash-1017 (35 tok/s decode) and Qwen3.8-Flash-Next (130 tok/s decode)

I'm sharing this here because this is working way better than it used to a few months ago with older models. I see people here posting slower speeds with much better hardware than mine and hopefully this information helps. I used gemini-flash to help me setup everything and I asked it to write up the sequence of steps to get me here - happy to provide more details if there's anything missing. Since I use gentoo some tools/commands might be unfamiliar to you. My 6 year old hardware: OS & CPU Gentoo Linux running on an AMD Ryzen Threadripper 3990X (64 cores, 128 threads). System Memory: 128 GB…

热度 50.5 / 100；排名与评分保留该期记录。