On May 7, Apple announced its latest homegrown computer chip, M4, which has been integrated into the new iPad Pro as its debut platform. The M4 boasts Apple's fastest neural engine ever, capable of performing up to 380 trillion operations per second, said to exceed the neural processing units of currently available AI PCs.
Apple says the Neural Engine and next-generation machine learning accelerators in the CPU, high-performance GPU, and high-bandwidth unified memory make the M4 an extremely powerful AI chip.
Internally, M4 is made up of 28 billion transistors, slightly more than M3. As for the process node, the chip is built on second-generation 3nm technology and acts as a system-on-chip (SoC), further increasing the efficiency of Apple chips.
According to reports, the M4 utilizes second-generation 3nm technology along the N3E process previously introduced by TSMC. According to TSMC, the density of N3E is not as high as N3B, but it offers better performance and power characteristics.
In terms of core architecture, the new CPU in the M4 chip features up to 10 cores, consisting of 4 performance cores and 6 efficiency cores, an increase of 2 efficiency cores compared to M3.
The new 10-core GPU is built on the next-generation GPU architecture introduced in M3 and brings dynamic caching, hardware-accelerated ray tracing, and hardware-accelerated mesh shading to iPad for the first time. M4 significantly improves professional rendering performance in applications like Octane, now 4x faster than M2.
Compared to the previous generation iPad Pro's powerful M2, the M4 offers a 1.5x improvement in CPU performance. Whether you're processing complex orchestral files in Logic Pro or adding demanding effects to 4K video in LumaFusion, M4 can improve performance across your professional workflow.
As for memory, the M4 chip adopts faster LPDDR5X, delivering a unified memory bandwidth of 120GB/s. LPDDR5X is a mid-term update to the LPDDR5 standard, offering higher memory clock speeds of up to 6400 MT/s. Currently, LPDDR5X speeds can reach up to 8533 MT/s, while M4's memory clock speeds only reach around 7700 MT/s.
According to industry data, the Apple M3 comes with up to 24GB of memory, but there is no data to indicate whether Apple will work on expanding the memory. New iPad Pro models will come with 8GB or 16GB of DRAM, depending on the specific model.
The new neural network engine integrated into the M4 chip has 16 cores and can perform 380 trillion operations per second. That's 60 times faster than the first neural network engine on the Apple A11 Bionic chip.
Additionally, the M4 chip employs an innovative display engine designed with cutting-edge technology to combine light from two OLED panels to create a state-of-the-art display with incredible precision, color accuracy, and an Ultra Retina XDR display. , achieving uniform brightness.
Johnny Sluge, Apple's senior vice president of hardware technology, said the M4's efficient performance and innovative display engine enable the iPad Pro's slim design and breakthrough display. Fundamental improvements to the CPU, GPU, Neural Engine, and memory systems make M4 ideal for modern AI-driven applications. Overall, this new chip makes iPad Pro the most powerful device of its kind.
- 2024 will be the first year of the AI PC era
Currently, AI is emerging as a superstar around the world. Apart from markets such as servers, the consumer market is embracing the new opportunity of AI PCs.
Previously, TrendForce predicted that edge AI applications will significantly expand in 2024, leveraging the foundation laid by AI servers and moving into AI PCs and other terminal devices. Demanding edge AI applications return to the AI PC, distributing the load on AI servers and expanding the potential for scale of AI usage. However, the definition of AI PC remains unclear.
According to Apple, the M4's Neural Engine is Apple's most powerful Neural Engine to date, outperforming any AI PC neural processing unit available today. Tim Millet, Apple's vice president of platform architecture, said the M4 will provide the same performance as he M2 while consuming half the power. Compared to next-generation PC chips in various lightweight laptops, M4 delivers the same performance while consuming only 1/4 the power.
Meanwhile, frequent developments by other major companies suggest that competition in the AI PC field is becoming more and more intense, and the industry also has high expectations for AI PC. Microsoft deemed his 2024 the “Year of the AI PC.” Based on PC brand manufacturers' estimated product launch schedules, Microsoft predicts that half of commercial computers will be AI PCs by 2026.
Intel has previously emphasized that AI PCs will be a turning point in the revival of the PC industry. AI PCs will play a key role in the industry highlights of 2024. Intel's Pat Gelsinger previously said at a conference that customers continue to add processor orders to his Intel, driven by demand for AI PCs and his Windows update cycle. Therefore, Intel's AI PC CPU shipments in 2024 are expected to exceed the initial target of 40 million units.
TrendForce expects the AI PC's computing power to meet Microsoft's benchmark of 40 TOPS. New products that meet this criteria are expected to ship at the end of 2024, and significant growth is expected in 2025, especially after Intel releases his Lunar Lake CPUs by the end of 2024.
Currently, the AI PC market is driven by two major factors. First, the demand for terminal applications, which is primarily dominated by Microsoft through its Windows OS and Office suite, is a key factor. Microsoft plans to integrate Copilot into the next generation of Windows and make Copilot a basic requirement for AI PCs.
Second, as a leading CPU manufacturer, Intel advocates AI PCs that combine CPU, GPU, and NPU architectures to enable a variety of terminal AI applications.
read more
(Photo courtesy of Apple)
Please note that this article quotes information from WeChat account DRAMeXchange.
Next article
[News] US imposes further sanctions, revokes Intel and Qualcomm's licenses to supply chips to Huawei
