Primate Labs today announced the release of Geekbench 7, the latest version of the popular CPU and GPU performance benchmarking suite.
There is a redesigned multi-core benchmark in Geekbench 7 intended to better represent how applications operate in the real world. A workload runs multithreaded only if the task it models runs multithreaded in real applications.
Primate Labs says users can expect a multi-core score that is a more relevant measure of a device’s performance during real-world work.
Alongside updates to multi-core testing, a revamped GPU benchmark focuses on machine learning and content creation to reflect how GPUs are used today. The GPU benchmark features tests that track faces and apply filter effects in real-time to video, enhance images using machine learning, and blur backgrounds in video conferencing streams. New GPU image editing and synthesis tests are available, such as RAW image processing, LUT-based video color grading, path tracing, and fluid simulation.
Nvidia’s CUDA is once again a supported API alongside OpenCL, Vulkan, and Metal.
The new version of Geekbench uses more complex data sets to better reflect how people use their computers today. File compression and PDF viewer workloads have more variation, and development and image processing tests have more assets and different image formats.
Geekbench 7 adds media workloads to test audio and video encoding, decoding, and processing. Whisper, AV1, and Opus speech recognition tests mimic video conferencing, screen sharing, and content consumption workloads.
There is also a Game Physics workload that uses the Jolt Physics engine used by popular video games. Primate Labs has added an expanded Photo Editor test with a richer set of real-world edits, and the Photo Library workload supports importing and processing image formats like JPEG XL and DNG.
Geekbench 7 is available for iOS, macOS and other platforms. The app remains free for personal use and the Pro version is 20% off the usual price of $99 until August 6.
