Nvidia RTX 3090 vs Intel Arc Pro B70 llama.cpp Benchmarks

Just sharing the results from experimenting with the B70 on my setup.... These results compare three llama.cpp execution paths on the same machine: RTX 3090 (Vulkan) on NixOS host, using main llama.cpp repo (compiled on 4/21/2026) Arc Pro B70 (Vulkan) on NixOS host, using main llama.cpp repo (compiled on 4/21/2026) Arc Pro B70 (SYCL) inside an Ubuntu 24.04 Docker container, using a separate SYCL-enabled llama-bench build from the aicss-genai/llama.cpp fork Prompt processing (pp512)