maeddesg/vulkanforge

LLM inference engine for AMD RDNA4 — Rust + Vulkan compute shaders, gguf & native FP8.

3 best points · 0 comments · 2 Hacker News posts · 24 GitHub stars · View on GitHub

Hacker News discussions

• VulkanForge – 14 MB Vulkan LLM engine that runs native FP8 models on AMD (Rust) (3 points, 0 comments, by maeddesg, 2026-05-03)
• Why Gemma-4 26B MoE works in HuggingFace but breaks in prod inference engines (1 points, 0 comments, by maeddesg, 2026-05-15)