Tech News | AI Benchmark Test Geekbench AI Launched to Evaluate Device Performance in Machine Learning
Primate Labs, the creators of the well-known benchmarking suite Geekbench, have rolled out a pioneering tool to assess machine learning capabilities called Geekbench AI. This cutting-edge benchmark is now available on iOS, Android, Windows, macOS, and Linux, representing a significant step in evaluating how devices handle real-world AI tasks.
Initially named Geekbench ML during its preview stage, Geekbench AI is tailored to provide a comprehensive assessment of device performance in the realm of artificial intelligence. The tool focuses on evaluating the efficiency and accuracy of CPUs, GPUs, and NPUs by measuring their ability to execute machine-learning tasks.
Geekbench AI generates three distinct scores for each test – single precision, half-precision, and quantized – providing insights not only into the speed at which a device can process AI workloads but also its precision. Moreover, the benchmark includes efficiency comparisons over time, offering users a detailed overview of their device’s AI capabilities.
Supporting a wide range of AI frameworks, Geekbench AI is versatile across various platforms. It integrates CoreML for macOS and iOS, OpenVINO for Windows and Linux, QNN for Snapdragon-powered Arm PCs, and several vendor-specific frameworks for Android devices. Each test undergoes a minimum of five iterations to ensure devices are evaluated at their peak performance levels.
Integrated into the Geekbench browser, this new tool allows for effortless cross-comparisons between different devices, empowering users to measure and contrast the AI processing power of their devices against others. With its ability to assess machine learning and artificial intelligence across diverse devices, Geekbench AI emerges as a valuable resource for consumers and developers alike.
