Qwen3.8-27B Serving Configs: DGX Spark vLLM and RTX 4090 llama.cpp
erdaltoprak · reddit · 2026-08-15
Reddit user erdaltoprak shares serving configurations for Qwen3.8-27B: one for Nvidia DGX Spark with vLLM (NVFP4, 262k context, MTP speculative decoding), and one for RTX 4090 with llama.cpp (GGUF Q4KM, 131k context, flash attention). Docker Compose files provided.
Related event: Community Shares Qwen3.8-27B Deployment on 32GB VRAM(4 posts)→
More from coding & agent
- Six-month prediction: hiring for generic agent harnesses built from scratch will cool — yi111 · 2026-10-16
- Omarchy ISO wouldn't boot on NVIDIA, so he had Claude patch it — indragie · 2026-10-16
- Dev uses Grok to clean up 3D printer firmware, catches it sneaking in nonsense code — yacineMTB · 2026-10-16
- A rude AI agent took 89% of a shared resource in 102-run multi-agent personality experiment — Brave-History-6502 · 2026-10-16
- Guardrails can make risky delegation work — but it takes a full-time team, says iandanforth — iandanforth · 2026-10-16
- What can you truly delegate? Only tasks with unambiguous, verifiable results — iandanforth · 2026-10-16