Geekbench, a well-known cross-platform benchmarking tool, recently launched its new AI testing tool - Geekbench AI. This tool is designed to evaluate the performance of devices under artificial intelligence-intensive workloads and provide users with a more comprehensive AI performance assessment by measuring the processing capabilities of CPU, GPU and NPU. This move marks Geekbench’s further expansion in the field of AI performance testing, providing developers and users with a more accurate reference for AI hardware performance.
Recently, the well-known benchmarking tool Geekbench released a new cross-platform tool specifically designed to evaluate the performance of devices under AI-intensive workloads. The new tool, called Geekbench AI, determines a device's ability to handle machine learning applications by measuring its CPU, GPU and NPU (neural processing unit).
Previously, Primate Labs, the development company of Geekbench, used the name Geekbench ML to develop the software and launched a preview version as early as 2021. In order to highlight its AI-focused features, it has been renamed Geekbench AI.
Geekbench AI works by exploring the capabilities of different hardware by evaluating their performance on AI-related tasks, including accuracy and speed. It also supports multiple frameworks, such as ONNX, CoreML, TensorFlow Lite and OpenVINO, so users can choose a suitable testing environment according to their needs. This tool will give three types of scores, namely full precision, half precision and quantization scores. At the same time, Primate Labs also stated that the score will be accompanied by an accuracy evaluation to help users understand the true extent of the workload output and the accuracy of the model's task execution.
As AI technology continues to advance, we need more time to observe how devices using on-device AI, such as Copilot Plus PCs and various new mobile phones, perform in real scenarios. In the past, we would focus on frame rate or load time, but now, perhaps we will pay more attention to the accuracy of predictive text, or the effect of a generative AI-powered image editor.
Users can download the Geekbench AI tool and try it on Windows, macOS, Linux, Android and iOS platforms to see how the device performs in AI processing.
Highlight:
- ?️ Geekbench AI is a new benchmarking tool focused on evaluating device performance on AI tasks.
- ⚙️ It provides accuracy and speed evaluation by measuring CPU, GPU and NPU, and supports multiple machine learning frameworks.
- The tool is now online and users can download and test the device’s AI processing capabilities on multiple platforms.
The launch of Geekbench AI provides a new standard for hardware performance evaluation in the AI field. Its cross-platform features and multi-framework support also make it widely applicable. In the future, with the continued development of AI technology, Geekbench AI may play an increasingly important role in AI performance evaluation, which is worth looking forward to.