oneapi

Run Ollama on AMD Ryzen 7 5800H CPU with integrated GPU AMD ATI Radeon Vega (gfx90c) with optimizations

Gentoo repository for VISA, DAQ, NI LabView, PicoScope, oneAPI and DKMS

Field-tested guide: multi-GPU vLLM tensor-parallel (TP=2/TP=4) on Intel Arc Pro B70 (Battlemage BMG-G31, Xe2) on Linux. Driver setup (xe force_probe=e223), bare-metal vLLM + oneAPI 2025.3, the compute-runtime multi-root USM + triton-xpu init_devices fixes, FP8/int4-AutoRound quant, root-cause error reports. AI-agent readable (AGENTS.md).

Setup scripts for the open source oneAPI

Docker Compose stack for serving a local, OpenAI-compatible LLM (vLLM on Intel XPU) on an Intel Arc Pro B60 GPU — reproducible config with operator and developer docs.

Serve DeepSeek-V4-Flash-0731 (162GB Q8_K_XL) fully RAM+VRAM resident on 2x Intel Arc Pro B70 via llama.cpp SYCL — 0.56 to 6.6 tok/s. Build recipe, serve config, system tuning, benchmark.