Optimal parameter settings for local deployment of Qwen3.8 Flash Next shared
Motor_Ad16 · reddit · 2026-08-28
A user shared specific parameter configurations for running Qwen3.8 Flash Next on a dual RTX 3090 and Threadripper Pro 3955 setup. Deployed via Llama CPP in an Ubuntu VM with flash-attn enabled and specific batch-size and ctx-size settings. Benchmarks show a text generation speed (TG) of 15-19 t/s and processing speed (PP) of 550 t/s. The author seeks advice on further optimization.
More from Infra
- Fluidstack hires Electrical Engineer for Modular R&D to build frontier AI infrastructure — MxMnr · 2026-10-16
- Jensen Huang joins BG2 podcast: does AI revenue justify capex, '27 bottlenecks — elonmusk · 2026-10-16
- Mistral's on-device infra rack revealed: DGX Spark, Mac Studio and more — gregmushen · 2026-10-16
- NVIDIA and Sweden's Wallenberg launch Sferical One, a shared AI supercomputer — nvidia · 2026-10-16
- Next infra frontier: self-evolving systems that adapt and self-verify — _shreya_s · 2026-10-16
- Xiaomi's MiMo RL dashboard reveals $3.47M spent on just 30 training steps — bycloud · 2026-10-16