Primate Labs is pleased to announce that Geekbench 7.1 is now available for download. This version introduces new functionality and important improvements:
-
CUDA support on Windows on ARM and Linux. Geekbench 7.1 adds CUDA support for Windows on ARM and for Linux. Users with NVIDIA GPUs can now run the GPU Benchmark using CUDA on these platforms.
-
TensorRT for RTX 1.6. Geekbench 7.1 upgrades TensorRT for RTX to version 1.6 and extends TensorRT for RTX support to Windows on ARM and Linux.
-
Vulkan cooperative matrix optimizer. Geekbench 7.1 adds an optimizer pass for the Vulkan cooperative matrix extension. This change improves performance in the GPU Benchmark’s machine learning workloads on GPUs that implement the cooperative matrix extension.
-
Improved compatibility with Geekbench 6. Geekbench 7.1 fixes conflicts that could occur when registering Geekbench 6 and Geekbench 7 on the same system.
-
Improved hardware identification on iOS. Geekbench 7.1 fixes issues that could cause Geekbench to misidentify the SoC on newer iOS devices.
Geekbench 7.1 CPU Benchmark scores are fully comparable with Geekbench 7.0 CPU Benchmark scores.
Geekbench 7.1 GPU Benchmark scores are not fully comparable with Geekbench 7.0 GPU Benchmark scores due to the CUDA and Vulkan improvements in Geekbench 7.1. We recommend comparing Geekbench 7.1 GPU Benchmark scores only with other Geekbench 7.1 GPU Benchmark scores.
Geekbench 7.1 is a recommended update for all Geekbench 7 users.