How about the performance of the A14 chip used on the iPhone 12?

The iPhone 12 that everyone is looking forward to is finally here. As expected, the entire iPhone 12 series are equipped with the A14 bionic chip. Although the A14 bionic chip was launched on the iPad Air 4 this year, the performance of the A14 bionic chip on the iPhone 12 is still very worth looking forward to. Let's use the article to understand what kind of experience the A14 bionic chip can bring us.


In terms of basic specifications, the A14 chip on the iPhone 12 uses an advanced 5nm process. Compared with the 7nm process of the previous generation A13 chip, the transistors on the A14 chip are more integrated. According to official data, 11.8 billion transistors are integrated on the A14 chip, while there are only 8.5 billion transistors on the A13 chip. Advanced process technology allows the A14 chip to achieve high performance while maintaining low power consumption.

The CPU of the A14 chip is a 6-core design (2 high-performance cores + 4 high-efficiency cores), and the GPU part is a 4-core design, which is consistent with the previous generation A13 chip. In terms of performance, according to the official information conversion, the performance of the CPU part of the A14 chip is improved by 16.6% over the previous generation, and the performance of the GPU part is improved by 8.3%. From these data, although the overall performance of the A14 chip has improved, the improvement is not large compared to the previous generation.

Theoretical performance

Because there is no real iPhone 12 at the moment, we can refer to the running score data exposed on the Internet and the performance of the iPad Air4 equipped with the A14 chip in the performance section.

Antutu once published a comparison chart of the running score of the suspected iPhone 12 Pro Max. The running score of this device was 572,333, while the average running score of the iPhone 11 Pro Max was 524,436. The running score of the CPU part of the A14 chip is about 16.7% higher than that of the A13, which is almost the same as the official announcement above. However, the improvement of the GPU part is different.

In addition, the iPad Air4 Geekbench 5 running scores exposed online also confirm the performance of the A14 chip.

Judging from the Geekbench 5 running score chart of the iPad Air4 exposed by foreign media, the A14 chip has a single-core CPU score of 1583 and a multi-core score of 4198. The iPhone 11 Pro Max equipped with the A13 chip has a single-core score of 1333 and a multi-core score of 3377. If you want to compare Android models, the Snapdragon 865 flagship processor on the Android platform scored 912 in single-core and 3319 in multi-core. From the point of view of scores alone, the A14 chip still has a large lead.

You can also see the core parameters of the A14 chip from the running graph. It can be seen that the frequency of the CPU part is close to 3.0GHz, and the cache part has increased compared to the A13 chip.

On the whole, the performance of the A14 chip has not been greatly improved, and it may take more into account the increased power consumption of 5G and give more consideration to battery life. Although the improvement is not large compared to the previous generation, there is no need to worry about performance issues. The A14 chip can still provide strong performance support.

The "League of Legends" game was also demonstrated at the press conference. With the support of the powerful performance of the A14 chip, it not only guarantees a smooth game experience, but also can see great screen details, even complex team battle screens can respond in time.

Neural network engine is greatly improved

The huge improvement of the A14 chip this time can be said to be in the neural network engine, which is what we often call NPU or machine learning capabilities. The A14 chip uses a new generation of 16-core neural network engine, which has been upgraded from the 8-core of the A13 chip to 16-core. Apple said that the neural network engine of the A14 chip can achieve 11 trillion operations per second, and the machine learning performance has been doubled in the past.

Although the increase in the computing power of the neural network engine is not as easy to detect as the increase in the performance of the CPU or GPU, it has actually been integrated into a lot of daily use experience, which is very important for the comprehensive experience of the entire phone.

Although there may not be obvious features, many daily usage scenarios use AI computing power to enhance the experience. Such as the most commonly used Face ID recognition function. When the iPhone is unlocked through Face ID recognition, it will also learn your face, so that the iPhone can master more of your facial features and unlock a higher success rate. In Face ID recognition training, these will be supported by the powerful neural network engine in the chip.

In terms of taking pictures, the A14 chip neural network engine can provide better photo effects. Through machine learning, iPhone 12 can understand the scene when taking pictures and make more detailed adjustments to different parts of the photo. Like the photo demonstrated above, the image is segmented through intelligent recognition of the sky, and the color, contrast, and noise of the screen are adjusted to highlight the stone part.

Similarly, there is the deep fusion function of taking pictures. The iPhone 12 will process the photos at the pixel level through the machine learning of the neural network engine, which can present unprecedented details and reduce noise.

In addition, the system will learn your daily software usage habits and provide a faster software opening experience. There are also translation functions provided by siri, etc., which require the support of powerful machine learning capabilities to present better results. The powerful neural network engine of the A14 chip can efficiently complete these computing requirements locally, reducing the computing pressure of the CPU and the need for networking, and providing a smoother and more efficient user experience.

In addition, the iPhone 12 Pro also includes a lidar scanner, with the machine learning capabilities of the A14 chip and the depth information framework of iOS 14. The iPhone 12 Pro can scan objects and rooms to create scenes with accurate depth information. Objects can be accurately placed in augmented reality, and rich application scenarios can be unlocked.


The A14 chip on the iPhone 12 is the same as the one previously launched on the iPad Air. Compared with the previous generation, the performance is not significantly improved, but it is still at the current leading level, so there is no need to worry about performance. With the blessing of more advanced processes, the A14 chip will have better performance in terms of power consumption, which will help improve the battery life of the iPhone 12.

The A14 chip has significantly improved the neural network engine. It can be seen that Apple attaches great importance to this part of the computing power. The upgrade of the neural network engine can bring greater imagination and unlock more application scenarios. The powerful neural network engine and the lidar of the iPhone 12 Pro can bring new experiences that are worth looking forward to. I believe this will also be some changing trends in the chip part in the future.

Although the A14 chip on the iPhone 12 does not have an amazing performance improvement, its overall performance is still quite excellent, and the actual performance on the iPhone 12 is also worth looking forward to.