We are excited to announce that Geekbench 4.1 is now available for download from the Geekbench website! This update brings Metal Compute Benchmarks to iOS and macOS, as well as changes to several of the CPU and Compute workloads.

Geekbench 4.1 includes the following changes:

  • Improve support for Ryzen processors

  • Fix memory leaks in OpenCL workloads

  • Fix a crash that could occur on Skylake-E processors

  • Fix crashes that could occur when running Compute Benchmarks on low-end GPUs.

CPU Workload Changes

Geekbench 4.1 includes the following changes to the CPU workloads:

  • Build Geekbench for Linux with Clang 3.9.

  • Build Geekbench for Android with Android NDK 13b.

  • Build Geekbench for iOS, macOS with Xcode 8.2.

  • Enabled AArch32 cryptography instructions in Android ARMv7 build.

  • Change Memory Latency workload to avoid cache hits on Cortex A72, A73.

  • Report Memory Latency workload performance in nanoseconds.

  • Add AVX512 implementations to FFT, GEMM workloads.

  • Disable SQLite cache statistics to improve multi-core scalability.

  • Disable LLVM runtime assertions to improve multi-core scalability.

Users can expect a 2% increase in single-core scores, and at least a 5% increase in multi-core scores. Note that the multi-core score increase depends on the number of processor cores – systems with more cores will see a larger increase in the multi-core score.

For more information about the CPU workloads, please refer to our CPU Workload whitepaper.

Compute Workload Changes

Geekbench 4.1 includes the following changes to the Compute workloads:

  • Consolidate optimization code for CUDA, Metal, and OpenCL workloads.

  • Convert Histogram Equalization, Sobel to use RenderScript intrinsics.

Users can expect a 35% increase in RenderScript scores, and a 5% increase in OpenCL and CUDA scores with Geekbench 4.1. These changes should also eliminate crashes on systems with low-end GPUs.

For more information about the Compute workloads, please refer to our Compute Workload whitepaper.

Comparing Geekbench 4 Scores

For the most accurate comparisons, we strongly recommend that users not compare Geekbench 4.0 scores with Geekbench 4.1 scores. While comparing overall scores may provide a rough approximation, comparing individual workload scores is strongly discouraged.

We do not expect to make any workload changes in future versions of Geekbench 4, so Geekbench 4.1.0 scores should be comparable against scores from all future releases of Geekbench 4.