Instead of just crunching digits of pi for an hour or doing 80 different versions of the same task, Geekbenchs tests measure tasks like how quickly a device can load an example website, render a PDF, add filters to photos, and process HDR. It's not likely that it will solve the overall problem of benchmarks (and how they aren't totally representative), but it further moves in the direction of highlighting that not everything is about raw CPU capability. New and Updated Real-World Tests Geekbench tests reflect how people use their devices. As such, new tests have also been added to Geekbench in order to better reflect how we really use our smartphones.įrom background blur tests (such as the ones you may use in video calls) to improved raytracing and AI object detection, there are a lot of improvements that will better reflect the actual holistic capabilities of your smartphone. Primate Labs, the developer of Geekbench, understands that machine learning and other heterogeneous computational workloads are the better metric for a smartphone's performance. In other words, all of these changes better reflect the smartphone and device usage of 2023, while being close to the workloads that apps process on a day-to-day basis. A larger library of images for import tests.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |