As the first benchmark results of Apple's A15 Bionic application processor appear on Geekbench, we can reveal some more information about the architecture of the new chip.
When Apple formally introduced its A15 Bionic system-on-chip (SoC) earlier this week, it said that it featured a 50% higher CPU performance and a 30% higher GPU performance when compared to an undisclosed competitor, but it never compared the it to its previous-generation A14 Bionic SoC. It did not take long before the first benchmark results of Apple's A15 Bionic hit the Internet though. As it turns out, the new SoC has a massively — 55% —higher performance in Geekbench 5 Metal graphics benchmark. For now, take these benchmarks with a grain of salt.
Apple A15 Bionic Performance in Geekbench 5 Metal
|A15 Bionic||14,216||iPhone 13 | iPhone 14.2||iOS 15.0|
|A14 Bionic||9,123||iPhone 12 Pro | iPhone 13.3||iOS 14.x|
|A13 Bionic||7,338||iPhone 11 Pro | iPhone 12.3||iOS 14.x|
Apple's A15 Bionic integrates two high-performance Avalanche CPU cores, four energy-efficient Blizzard CPU cores, a 4-cluster or 5-cluster GPU, and a 16-cluster NPU. Apple's new CPU cores can now work at up to 3.23 GHz, up from 2.99 GHz in case of last year's A14 Bionic. An 8% CPU frequency increase is something that was logical to expect from the new SoC as it is presumably made using TSMC's performance enhanced N5P process technology.
With GPU, the situation seems to be considerably more interesting. The A15 Bionic that goes into iPhone 13 (presumably model iPhone 14.2) and iPhone 13 mini (likely model iPhone 14.1) will feature four GPU clusters. By contrast, the version of the chip that is used for the iPhone 13 Pro (iPhone 14.3) and the iPhone 13 Pro Max (iPhone 14.4) has five GPU clusters enabled, hence the Pro versions will offer higher graphics performance. Apple's previous-generation A14 Bionic comes with four GPU clusters no matter in which device.
Primate Labs Geekbench 5 graphics benchmark is not the best measuring tool and certainly synthetic benchmarks do not always reflect performance in actual games. Yet, a massive 55% performance uplift that the iPhone 13 (non-Pro) has over an iPhone 12 Pro indicates that Apple uses a brand-new GPU architecture in its latest SoC. Since the new chip packs substantially more transistors (15 billion vs. 11.8 billion) than its predecessor, architectural improvements are certainly likely to expect.
While graphics performance is certainly important for smartphones, a new GPU architecture will certainly be welcome by Apple's next-generation PCs running new Apple Silicon SoCs. Meanwhile, we have no idea when those processors adopt this new GPU architecture.
At this point benchmark results obtained on an unreleased iPhone should be considered with a grain of salt since we do not know whether we are dealing with a final or pre-release version. Furthermore, testing conditions are also unclear. In short, while it looks like Apple's A15 Bionic has a faster GPU, wait for scores obtained on commercial hardware.