Geekbench 6.1 update increases single and multi-core results by up to 10%

Geekbench has recently released its latest iteration, in the form of the Geekbench 6.1 update. This new version brings significant improvements to the performance of both single-core and multi-core CPUs, with potential increases of up to 5% and 10%, respectively, depending on the system configuration.

One notable addition in Geekbench 6.1 is the introduction of the AV512-FP16 instruction, specifically designed for image processing. This instruction enables more efficient handling of image-related tasks. Additionally, the implementation of the SVE (Scalable Vector Extension) instruction enhances the benchmark's machine learning capabilities, further boosting overall performance.

Moreover, notable changes have been made to the usage of background blur and horizon detection, particularly benefiting high-end systems equipped with 12 or more cores. This improvement will likely be more noticeable on HEDT systems and servers, as they tend to have CPUs with a higher core count.

Due to these and other changes (Clang 16, fixed-point math, increased workload gap), Geekbench scores in version 6.1 may differ from those in version 6.0. Consequently, users should be aware that the results may not be directly comparable between the two versions. The official CPU ranking results should now incorporate a margin of error of up to 5% in single-core and up to 10% in multi-core. As the benchmark data for version 6.1 continues to accumulate over the next few weeks, the rankings will become more representative.

KitGuru says: Have you run Geekbench 6 on your current system? Did your score increase when running Geekbench 6.1?

Become a Patron!