Even those who aren't into 3D modeling or rendering have probably heard about Cinebench before. That's because it is often used by reviewers to benchmark Intel and AMD processors to test their multi-threaded might. Cinebench is a rendering benchmark based on Cinema 4D and is meant to stress all available cores and threads in the CPU.
Maxon, the creators of the benchmark has released an update to Cinebench today. The new version called Cinebench 2024 brings many major upgrades over its predecessor, Cinebench R23, as it now supports measuring GPU performance as well. In the past, Cinebench did have an in-built GPU benchmark, an OpenGL car chase sequence. However, that was dropped as it did not often match real-world performance. Cinebench 2024 however brings an integrated RedShift rendering engine which renders the same scene as the CPU benchmark, making it possible to compare CPU and GPU rendering times.
Aside from GPU support, Cinebench 2024 also comes with Windows on Arm support now as Qualcomm chips will run Cinebench thanks to a new ARM64 variant of the software. Aside from that, Apple silicon, i.e., M1 and M2 series of processors has also gained support with this latest edition.
The press release says:
Wider Platform Compatibility
Cinebench 2024 is designed to accommodate a broader range of hardware configurations. It seamlessly supports x86/64 architecture (Intel/AMD) on Windows and macOS, as well as Arm64 architecture to extend its reach to Apple silicon on macOS and Snapdragon® compute silicon on Windows, ensuring compatibility with the latest advancements in hardware technology. Redshift GPU performance can be evaluated on systems with compatible Nvidia, AMD and Apple graphics processors.
The press release also talks about some of the core enhancements it made to the codebase so that the new version of Cinebench is more intense. The memory footprint has been greatly increased to simulate memory-intense workloads. Aside from that, the compute intensity has also been raised six-fold, says Maxon, and they probably aren't lying. I tested it on my Ryzen 7 5700G and it took far longer than I anticipated as the Cinebench R23 run generally completes within a couple of minutes. This one though took close to five minutes.
Maxon writes:
Under-the-Hood Enhancements
Beyond the surface, Cinebench 2024 brings forth a host of performance-enhancing features. With a threefold increase in memory footprint compared to Cinebench R23, the software caters to the memory-intensive demands of modern projects. Moreover, a six-fold rise in computational effort and utilization of newer instruction sets ensures a benchmark that resonates with the complexity and sophistication of contemporary creative projects.
Here are the screenshots of my 5700G and Radeon 6800 XT running Cinebench 2024:
 |
 |
As you may notice, my 5700G is performing especially poorly in the test and that is because it's undervolted so that my A520 motherboard VRM, which lacks heatsinks, does not overheat. The 6800 XT though is at stock clocks and performing as well as it should.
3 Comments - Add comment