Apple held its 2019 autumn conference at 1 a.m. Beijing time on the 11th. Every autumn conference focuses on the iPhone, and this time the upgrade of the iPhone 11 is quite significant compared to the X to XS generation. The A13 chip has regular hardware upgrades, lens upgrades, the addition of night mode, and increased battery capacity, with the pro version standard equipped with 18W fast charging. The screen has been upgraded with double the contrast, and the entire series supports dual SIM dual standby. It seems that this generation of iPhone has filled in all the gaps except for supporting 5G, keeping pace with Android.
Calculating back, it has been 12 years since the first iPhone was released. The changes in appearance from the first generation iPhone to the current 11 are undoubtedly huge, and the internal hardware has also evolved significantly. In these 12 years, what Apple is proud of, besides the iOS system, is its self-developed A-series processors.
Starting with the A4 processor launched with the iPhone 4, Apple officially embarked on the path of self-developed chips. From A4 to now A13, Apple has had a total of 10 generations of A-series processors, and today, Jiemi Network will discuss the history of the A series from ancient to modern times.
From A4 to A13, Apple has not just been squeezing out incremental upgrades; on the contrary, several processors have had epoch-making significance.
The A4, as Apple’s maiden work, uses a 45nm process 800MHz ARM Cortex-A8 single-core processor, with a GPU of PowerVR SGX 535 and an L2 cache of 640KB. At the same frequency, its performance is better than Samsung‘s S5PC110, but its core structure is very similar to the Samsung processors used before. Strictly speaking, this chip cannot be considered a truly independent research and development achievement by Apple.
The A5, released with Steve Jobs‘s last work, the iPhone 4S, is the first dual-core processor chip designed by Apple. The changes in A5 go far beyond just adding a core; its processor architecture has also been upgraded from the previous Cortex-A8 architecture to the more powerful Cortex-A9. This new architecture, which evolved from the previous Cortex-A8 platform, has higher computing power and lower power consumption. The GPU part is equipped with PowerVR SGX543+, with significant improvements in graphics performance. It is said that the CPU performance is twice that of the first iPad, and the GPU is nine times that of the first iPad. The A5X is an enhanced version of its performance, featuring a quad-core graphics processor for the third-generation iPad, with graphics processing capabilities twice that of the iPad 2.
The A6, designed by Apple’s subsidiary Intrinsity and manufactured by Samsung, uses a unique architecture design, with performance between Cortex-A9 and Cortex-A5, based on a 32nm process. It can dynamically adjust the CPU voltage/frequency characteristics, and the integrated GPU is a three-core PowerVR SGX 543MP3 graphics processing unit, with performance more than twice that of the A5. Similar to the A5, the A6X is specially designed for the iPad, with increased CPU frequency and a GPU upgraded to SGX554MP4, featuring four cores.
When the A6 was released, Apple began to reach a peak of performance. The iPhone 5, equipped with the A6, outperformed many Android flagship devices with four-core processors in benchmark scores, proving to the world that the performance of a processor cannot be judged solely by the number of cores.
The A7 marked a new height for Apple and opened the door to the 64-bit era for mobile processors, undoubtedly a chip of great epoch-making significance.
The A7 adopts a new 64-bit design using the Arm-v8 64-bit instruction set and Apple’s own Cyclone architecture, manufactured using a 28nm process with a frequency of 1.3GHz. The performance of the A7 processor is twice as fast as the A6 on the iPhone 5, 40 times that of the processor used in the first-generation iPhone, and the graphics capability is 56 times that of the first generation. Moreover, starting from the A7, Apple equipped its processors with low-power co-processors of the M series, specifically responsible for processing various sensors data on the phone while maintaining extremely low power consumption.
As the first mobile 64-bit processor in history, it caused quite a stir. While other Android manufacturers were desperately increasing CPU core counts, Apple pointed out a new direction for future CPU development, which led to some criticism of the A7 at its launch, with many believing it was just a gimmick. However, the facts proved that the 64-bit architecture was indeed a very forward-thinking idea from Apple. At that time, people had already started editing videos and photos on their phones, and the performance of 32-bit processors was nearing its limits. Apple decisively entered the market early, allowing it to maintain an edge over Android manufacturers. Even today, while 8-core processors are standard in thousand-yuan machines, Apple still uses a 6-core design.
Subsequently, the A8, A9, and A10 further increased core counts and process technology, with processor and GPU performance climbing steadily. Notably, the A8X processor equipped for the iPad Air 2 was the first triple-core processor in the mobile industry, with incredibly powerful performance. The A8X integrated about 3 billion transistors, one and a half times more than the A8, and three times that of the previous A7, with a frequency increased to 1.5GHz. Its single-threaded performance was nearly 12% higher than that of the A8, and multi-threaded performance increased by a full 55%. The third core’s performance was perfectly released, and Apple first equipped the iPad Air 2 with 2GB of RAM. It must be said that the performance gap between this generation of iPhone and iPad was very significant.
Of course, there was also the A10X. When the A10 Fusion reached four cores, the A10X directly jumped to six cores with a 3+3 design, plus a 12-core GPU. Apple claimed it could easily edit 4K videos and render detailed 3D models, creating and marking complex documents. Looking at which device the A10X was first launched on—the iPad Pro—it is clear why it had such powerful performance. It was not just a superficial claim to be the strongest mobile CPU in history. Even now, the GPU performance of the A10X still dominates the charts, and its CPU performance is not much weaker than that of the A11 and A12.
Fast forward to 2017, a year that marked the beginning of a new era for both Apple’s A-series processors and the iPhone. The iPhone 8 beautifully concluded the previous iPhone design language, while the iPhone X ushered in the era of full screens and Face ID.
That year, the A11 Bionic was a performance monster. The A11 adopted TSMC’s most advanced 10nm process at the time, with 4.3 billion transistors and a six-core design. The performance of the large cores improved by 25% compared to the A10, and the four small cores improved by 70%, with multi-performance processing increased by 75%. The GPU, being Apple’s self-developed three-core GPU, improved by 30% compared to the A10, while power consumption was reduced by 50%. Additionally, the A11 was the first to feature a neural network engine with a dual-core design, capable of performing 600 billion calculations per second, primarily for machine learning tasks, enabling it to recognize people, places, and objects, with the most typical application being its first launch of Face ID and its derivative animated expressions.
The deeper significance of the A11 for Apple lies in the fact that it was the most self-sufficient generation of A-series processors, including self-developed CPU, self-developed GPU, self-developed ISP, self-developed decoder, and so on, including the neural network engine. From that year on, almost all mobile phone manufacturers began to highlight AI computing capabilities as the main focus of their marketing presentations, marking the beginning of the AI era.
By last year, with the release of the A12, Apple further improved performance and energy efficiency, thanks to TSMC’s latest 7nm process. The A12 adopted Apple’s self-developed Fusion architecture, and according to Apple’s data, the two large cores improved by 15%, while the four small cores improved power consumption by 50%. The GPU adopted a new generation of self-developed chips, with the core count increased to four, claiming a performance improvement of 50%, while the most significant upgrade was the neural network engine, which increased from dual-core to eight-core, capable of achieving 500 trillion calculations per second.
Finally, the A13 chip released early this morning boasts two major highlights: machine learning capabilities and low power consumption. The A13 Bionic uses TSMC’s upgraded 7nm process, integrating 8.5 billion transistors. The CPU has two high-performance cores, providing a 20% speed boost while reducing power consumption by 30%. It also has four efficiency cores, with the same speed increase and a 40% reduction in power consumption. The CPU can perform one trillion calculations per second. The GPU is designed with four cores, achieving a 20% speed increase and a 40% reduction in power consumption. Additionally, the A13 features an eight-core neural computing engine, with a performance increase of 20% and a power consumption reduction of 15%.
In fact, in recent years, not only Apple but other chip manufacturers have also changed from initially pursuing performance excellence to pursuing extreme energy efficiency, thanks to the advanced process technology of wafer foundries. Powerful performance inevitably increases energy consumption, and today, unless pursuing high-end gaming or multitasking, the performance of mobile devices is often excessive. A thousand-yuan machine’s performance is sufficient for daily needs, making battery life the most important aspect for mobile devices. Fortunately, Apple has also realized this this year, not only improving the chip but also increasing battery capacity and equipping it with an 18W charger.
From this chart compiled by Bloomberg, we can clearly understand several major events in the history of Apple’s A series. From A4 to A11, it has only been 8 years, but the performance of the A-series processors has risen too quickly. We marvel at Apple’s self-research capabilities and forward-thinking vision; it truly deserves to be called a great enterprise.
However, expressing this solely in words undoubtedly lacks intuitiveness. Therefore, I have collected the GeekBench scores of various generations of A-series chips into a table to visually display Apple’s formidable chip performance. (Note: Due to different versions of GeekBench, there will be some algorithm differences, so the scores of earlier chips compared to the current newer chips have certain algorithmic discrepancies. The score data is for reference only, and the GeekBench database contains a vast amount of data, with the score being just one of many, not the average.)
From the chart, we can clearly see several important nodes of the A-series processors: A6, A7, A8X, A10X, A11, each representing a leap in Apple’s chip strength. Based on the 20% improvement of A13 compared to A12, the A13 has the capability of 70 times that of the A4 processor level, and this is merely within 10 years.
People often say that today’s Apple no longer has the same dominance and influence as in the era of Steve Jobs. In terms of the iPhone, this is indeed true; the rise of Android manufacturers and the emergence of Chinese domestic mobile phone manufacturers in the global market have had a significant impact on Apple. However, it cannot be denied that Apple remains a great enterprise, at least it once changed the world.
It is not easy to leave a significant mark in history, and how many can achieve lasting fame?
Source: C114
* Disclaimer: The sharing and reprinting of the mobile phone market is solely to convey a different perspective and does not represent Mobile Market Sharing‘s agreement or support for this viewpoint. If there are any objections, please feel free to contact Mobile Market Sharing.