Strix Halo (8060S / gfx1151), Qwen-3.8-27B @ Q8 and Q6 UD v3, up to 256K ctx, llama.cpp, DFlash2, vision, real workloads quality and steady performances, optimized recipes, ...

Hi fellows fully-local halos, after manually following existing guides, I decided to build an LLM API endpoint installation and optimization guide that works even when autonomously followed by my pi agent, so I can install/experiment/reinstall easily and without babysitting. Q8 is my default citizen, options for Q6 and Q5. Recipes: Quality (Q8), Balanced (Q6), Speed (Q5), Vision (Q8). All with Unsloth Dynamic Quants 3.0, DFlash2 (except vision). Scripts for download the right LLMs, interactive testing, systemd --user install, adaptive quality and performances optimization. Repo: github.com/PieBru/Qwen-3.8-27B_Strix-Halo_gfx1151 Humans architected, verified, sealed. AI assistants built and wrote all the delivered stuff, built with pi and Qwen-3.8-27B. Piero P.S.: constructive proposals and PR are welcome. P.P.S.: no speed races, please. IMO speed is useful, but quality is fundamental - one subtle bug fewer or a better codebase always pays for itself in wall-time gained.