AMD's Loom + Gemini agent tunes kernels in 4h: 1.88x decode, up to 5.9x prefill on Strix Halo
sa1sr1 · reddit · 2026-11-21
An AMD engineer (on the Loom team, no early weight access) used Gemini 4 Neon with the pi.dev coding agent and the loom-skills playbook to optimize Inkling-2-Mini kernels on Strix Halo in 4.3 hours / 490 tool calls. Results vs best llama.cpp (ROCm/Vulkan): decode 44.8 vs 23.8 tok/s (1.88x, 80% of bandwidth roofline), prefill up to 5.9x at 18k context, 4-stream aggregate throughput 1.86x. Correctness gates (37k-token goldens, tool-calling, concurrency) passed before speed counted. Full install and serve commands included.
More from coding & agent
- mnemos: a self-curating agent memory plugin for Hermes that fades unused memories — RileyRalmuto · 2026-11-21
- Study: evolutionary LLM search found valid solutions but picked invalid programs in 2 of 3 runs — Unusual-Spot-5846 · 2026-11-21
- Dev routes tool calls through a 421M model at 12ms to catch risky agent actions — looselyhuman · 2026-11-21
- Screen-share + voice via MCP into Claude: a faster UI iteration workflow — Different_Scholar_74 · 2026-11-21
- Tencent & KCL release GroupAssistBench: group-chat agents recall memories but fail to use them — 量子位 · 2026-11-21
- Open-source E2E-encrypted MCP server lets your phone message a local AI agent — uriwa · 2026-11-21