MLPerf Client as MLCommons’ machine language inferencing benchmark for client form factors / PCs now has a Linux build. MLPerf Client 1.5 was released yesterday with an experimental Linux build but for now at least is not nearly as full-featured as this AI benchmark on Windows and macOS.
MLPerf Client is designed to evaluate large language models (LLMs) and other AI workloads on hardware from laptops and desktops to workstations, compared to MLPerf proper running on GPU-accelerated AI servers.
MLPerf Client on Windows x64 supports a range of GPUs, NPUs, and CPUs with different execution providers / APIs. For Apple iOS/macOS there is also accelerator support via Metal and MLX. But for the initial Linux build it just supports OpenVINO… For CPU execution or the Intel NPU if you get that working well and the LLM isn’t too big.
So for now at least the MLPerf Client for Linux isn’t too interesting without any other targets being supported besides OpenVINO.
MLPerf Client for Linux is command-line only (CLI) without any GUI like found on other platforms.
Those wanting to check out the MLPerf Client 1.5 release or learn more about the other changes can do so via GitHub.
