The M4 chip is coming, the AI ability is bursting, can Apple surpass it?
Apple unveiled the new M4 chip at its launch event in May this year, and the chip, which is made on TSMC's 3nm process, is seen as an important step in Apple's AI field. The M4 chip not only has a significant improvement in CPU and GPU performance, but also its powerful AI computing power, which is claimed to be more powerful than any AI PC chip today.
The AI strength of the M4 chip is mainly reflected in its neural network engine. The IP module, designed to accelerate AI workloads, has a computing power of up to 38 trillion operations per second, which is a staggering 60 times faster than Apple's first Neural Engine. With this powerful AI accelerator, coupled with a high-performance CPU, GPU, and unified memory with higher bandwidth, the M4 chip is positioned by Apple as an extremely powerful AI chip.
At the press conference, Apple showcased the M4's amazing performance in AI acceleration. The new iPad Pro with the M4 chip can instantly separate 4K video from the background in Final Cut Pro with a single tap. Users can even simply listen to someone play the piano and the StaffPad app generates scores in real-time. These seemingly magical AI functions are derived from the powerful neural network acceleration capabilities of the M4 chip.
Apple emphasizes that AI inference on the M4 chip can be done efficiently and privately, with minimal impact on application memory, responsiveness, and battery life. This means that future Apple devices will be able to provide users with a smarter and smoother AI experience, without the problems of power consumption and heat generation like other manufacturers' AI devices.
In addition to AI, the M4 chip also has a lot of improvements in CPU and GPU performance. According to the data released by Apple, the CPU performance of the M4 is 50% faster than that of the M2, and the GPU performance is 4 times faster than that of the M2. This is mainly due to the fact that the M4 uses a CPU architecture with 4 large cores and 6 small cores, as well as a 10-core GPU. It is also believed that the performance improvement of the GPU of M4 is limited, which is jokingly called "squeezing toothpaste".
The huge improvement in AI performance of the M4 chip naturally makes people look forward to its comparison with competing products. According to the available information, the M4 will far surpass other mobile chips in the same period in terms of AI computing power.
Taking Qualcomm's flagship mobile chip Snapdragon 8 Gen2 as an example, its AI computing power is 24 trillion operations per second, while the M4's neural network engine computing power is as high as 38 trillion operations per second, a difference of an order of magnitude. Intel has announced that its future Ultra mobile chip AI computing power will reach 34 trillion times per second, but M4 will still lead.
The reason why the M4's AI performance is so strong is largely due to Apple's advantages in software and hardware synergy. Apple has not only added a dedicated AI accelerator at the hardware level, but also deeply optimized the macOS system at the software level to maximize the AI capabilities of the M4 chip.
For example, Apple has implanted a large number of AI functions in the macOS system, covering images, video, voice and other fields. When users use professional software such as Final Cut Pro, M4's AI acceleration capabilities will greatly improve work efficiency.
Apple has also invested a lot of resources in developing its own large models. M4 with Apple's self-developed large model, will bring users a smarter and more humanized experience. Users may be able to complete complex tasks with simple voice commands.
Apple's layout in the field of AI did not happen overnight. As early as the mid-2010s, Apple began to add AI elements at the chip and system level, such as neural network engines, machine learning accelerators, etc. After years of hard work, Apple is finally able to outperform its competitors in AI performance.
The arrival of the M4 chip will undoubtedly bring a new experience to the Mac product line. AI is changing the way humans interact with each other, and Mac, as one of Apple's core product lines, naturally needs to fully embrace the AI era.
M4 will greatly boost the computing power of your Mac. Whether it is graphics rendering, video editing, or scientific computing, etc., it will be accelerated by AI to get twice the result with half the effort. Users will have an unprecedented and efficient experience when using professional software.
M4 will allow Macs to interact more intelligently. Users may only need to speak voice commands to the Mac, and the computer can complete complex tasks such as document editing, image processing, etc., greatly improving work efficiency.
M4 will drive the further development of the Mac ecosystem. With its powerful AI capabilities, Mac will attract more developers, launch more innovative apps, and bring more choices to users.
The emergence of the M4 also means that the Mac product line will usher in a new upgrade cycle. Apple has been criticized for the slow pace of upgrades to the Mac line, but with the M4, the Mac may have a similar annual iteration rhythm as the iPhone in the future.
It remains to be seen whether the M4 chip will help the Mac regain its strength. The decline in Mac sales did not happen overnight, in addition to performance factors, there are many reasons such as ecology and pricing.
If the M4 can truly revolutionize the user experience and eventually drive a rebound in Mac sales, it will become a jewel in the history of Apple silicon. On the contrary, if the "AI light" of M4 is just a flash in the pan, then it will also be unable to escape the fate of "flash in the pan".
Although the M4 chip is full of tempting imagination, to achieve a complete revival of the Mac product line, I am afraid that Apple's joint efforts in ecology, marketing and other aspects are needed. It is expected that the M4 chip will bring new vitality to the Mac product line, and I also hope that Apple can make this vitality bloom for a long time.