Qwen3.8 27B
on NVIDIA RTX PRO 6000 Max-Q · 192,000 ctx
Summary
User reports decode speed rising from 18 tok/s to 58 tok/s with DFlash2 + XQA versus Flash2 + FA2 at 192K context.
Lossless BF16 weights and KV.
on NVIDIA RTX PRO 6000 Max-Q · 192,000 ctx