The widespread benchmarking significance Geekbench has launched a new cross-platform tool to guage the efficiency of units beneath AI-heavy workloads. Geekbench AI measures a tool’s CPU, GPU, and NPU (neural processing unit) to decide how neatly it will possibly deal with system finding out programs.
Geekbench developer Primate Labs has been running at the tool the use of the title Geekbench ML, which launched in preview in 2021 however shifted the title to AI for causes that appear clear. To discover how other {hardware} responds to other AI-related duties, it evaluates efficiency in response to each accuracy and pace, with assistance for various frameworks, together with ONNX, CoreML, TensorFlow Lite, and OpenVINO.
It delivers 3 ratings, complete precision, part precision, and quantized. Primate Labs says the ratings even have an accuracy dimension to guage how similar a workload’s outputs are to the reality, “or how accurately that model can do what it’s supposed to do.”
We’ll want extra era with units operating on-device AI such as Copilot Plus PCs and all the pristine telephones to peer how efficiency in real-world duties correlates with Geekbench AI’s numbers. Checking framerates or loading occasions is something — now we may well be checking the accuracy of predictive textual content, or what generative AI-enabled symbol essayist comes up with.
You’ll be able to download the instrument now to attempt it your self on Home windows, macOS, Linux, Android, and iOS.
