RTX 6000 四卡本地部署 V4.1 Flash：单流 340 tok/s NVFP4
用户改进推理配方后在 4 张 RTX 6000 上以 NVFP4 跑 DeepSeek V4.1 1017，单流达 340 tok/s，称此前 160 tok/s 因配置不全。
Modified version of @fraserpricee 's recipe is printing deepseek v4.1 1017 tokens at 340 tok/s single stream which is the fastest I've experienced on my own hardware for NVFP4. Turns out previous 160 tok/s results were largely because fraser's recipe was incomplete and didn't contain all configs needed. Fraser was helpful (and had a much lower boast/receipt ratio) in DM communication. 😅 The claimed 700+ tok/s is still largely based on measurement yardstick optimization. Wasn't easy getting the man to acknowledge that standardized measurement is just basic intellectual honesty, but the stack is real. Below C8 this is the fastest I'm aware of for serving ds41f on tp4 rtx6k. llm-inference-bench image below shows a modified version of BRRR. I'm calling it BRRR/B12X hybrid as it uses native B12X above 8 scheduled tokens which leads to some gains at higher concurrency. Recipe in 🧵
这条是引用着另一条写的：x.com/zenmagnets/status/2114768356339568829