When Apple introduced the iPhone 12 series earlier this week, the A14 Bionic inside it was called the fastest smartphone chip out there. Apple claimed a CPU and GPU performance superiority of up to 50% compared to the most powerful non-Apple chip in smartphones. Now, that’s a huge number, and assuming that the fire-breathing Snapdragon 865 was the fastest non-Apple chip the company has in mind, the A14 Bionic is truly going to set a new benchmark. Talking about benchmarks though, early synthetic benchmark figures of the A14 Bionic-powered iPhone 12 Pro are in, and the performance gain it brings is truly impressive.
First spotted by the folks over at Macrumors, the iPhone 12 Pro posts an average single-core score of around 1600, while the multi-core tally can be seen touching the 4,000 mark. Compare that to the iPhone 11 Pro, which averages around 1300 in single-core tests, while the multi-core score hovers around the 3300 mark. Now, if you calculate the difference, the A14 Bionic edges out the A13 Bionic inside the iPhone 11 Pro by a margin of 20% on both single-core and multi-core performance tests.
Macrumors claims that the performance gain offered by the A14 Bionic silicon even touches the 25% mark in some cases. And in case you’re wondering about abnormally low single-core tally on some Geekbench v5.0 tests of the iPhone 12 Pro models, Geekbench founder John Poole reportedly explained to MacRumors that it might be because “the devices could be in the process of being restored among other factors.” You can read more about what the A14 Bionic chip has to offer here.
However, it is still quite impressive that the A14 Bionic delivers up to 20% performance boost over last year’s A13 Bionic, a chipset that continues to give Qualcomm’s flagship chipsets a run for their money to this day. The A14 Bionic inside the iPhone 12 Pro is also the first smartphone chip based on the 5nm process. It has a hexa-core design and comes equipped with a new 16-core Neural Engine that is claimed to be almost twice as fast compared to its predecessor and can perform 11 trillion operations per second. It also comes with second-generation machine learning accelerators that are touted to be 10 times faster at performing machine learning calculations.