Geekbench 3.1.0

Geekbench 3.1.0 is now available for download. Geekbench 3.1.0 features the following changes:

  • Introduced 64-bit benchmarks on iOS.
  • Added ARM cryptography instruction implementations of AES and SHA-1.
  • Re-enabled standard optimizations for the STREAM Copy workload.
  • Reduced the memory footprint for the BlackScholes and STREAM workloads.
  • Added support for older Linux distributions.
  • Fixed a crash that could occur on dual-processor systems running Windows.
  • Fixed a crash that could occur on 32-core systems running Windows.

This release makes some changes to the workloads used to measure performance. While we try and keep benchmark results comparable across minor updates, some of these changes might result in small differences between 3.0 results and 3.1 results. In particular, I'd like to point out the changes that might have the biggest impact on performance:

  • Geekbench 3.1 now includes 64-bit benchmarks on iOS. Expect higher performance with Geekbench 3.1 on the new iPhone 5s than with Geekbench 3.0.

  • Geekbench 3.1 uses the new ARM cryptography instructions in the AES and SHA-1 workloads. These new instructions will significantly boost the performance of these two workloads on processors with the new ARM cryptography instructions. Currently only the Apple A7 supports these instructions, but we expect that to change over the coming months. Note that Geekbench 3 already uses the equivalent Intel cryptography instructions for the AES workload.

  • Geekbench 3.0 disabled a common optimization for the STREAM Copy workload that would allow the compiler to use highly-optimized library functions. Unfortunately we were only able to disable this optimization on some platforms and not others. This resulted in significant differences across platforms, even when running on the same hardware. Geekbench 3.1 re-enables this optimization, making STREAM Copy performance more comparable across platforms.

Geekbench 3.1.0 is a free update for all Geekbench 3 users.


 
John Poole is the founder of Primate Labs and lives in Toronto, Ontario with his wife Deborah. You can find John on Twitter or .