informatico-madrid/blackwell-linux-infra-optimizer

Optimized vLLM deployment for NVIDIA Blackwell (RTX 5090) on Linux Kernel 6.14. Resolves SM_120 kernel incompatibilities, P2P deadlocks, and memory fragmentation for high-performance LLM inference.

2 best points · 0 comments · 1 Hacker News post · 6 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: Optimized vLLM for Blackwell (SM_120) – 59T/S on DeepSeek-R1 (2 points, 0 comments, by malka666, 2026-01-17)