A system on a chip or SoC is a specific type of an integrated circuit comprising most or all of the components of a computer or other electronic systems. For comparison, a circuit board or motherboard used in traditional electronic system architecture has separate hardware components with different specialized functions. These components connect to a central interfacing circuit board, thus making them detachable and replaceable. An SoC fundamentally integrates these components in a single integrated circuit or chip.
In devices such as personal computers, including desktops and laptops, smartphones, tablet computers, and wearables, a system on a chip would typically include microcontrollers, a central processing unit or several processor cores, a dedicated graphics processing unit, other coprocessors such as artificial intelligence or machine learning engines, and radio modems such as Bluetooth, Wi-Fi, and cellular network radios, among others.
Integration defines the core advantages of an SoC. The commercialization of this technology has played a central role in the introduction and further developments on mobile consumer electronics technologies, particularly the introduction and subsequent progress in smartphones, tablets, laptop computers, smart devices, and wearable products. However, integrating different components with different functionalities in a single chip has drawbacks or disadvantages.
Pros of SoC: Advantages of System on a Chip
Suitability for Mobile Consumer Electronic Devices
A considerably smaller footprint is one of the advantages of a system on a chip. Modern SoCs have a physical dimension that ranges from 10 to 15 square millimeters. On the other hand, a motherboard has a larger real estate because it needs to accommodate separate hardware components of different shapes and sizes.
The smaller physical dimension of SoCs made them suitable for portable or mobile consumer electronic devices. As mentioned, the introduction, as well as the further advancements in the design and production of these chips paved the way for the introduction and continuous developments in smartphones and tablet computers.
A separate CPU attachable to a motherboard can have the same size or even larger than a system on a chip. Hence, using traditional electronic system architecture on mobile devices is impractical, especially if the goal is to make these products as portable as possible while making enough space for batteries and other hardware components.
The larger hardware components of a motherboard require more energy. Mobile devices depend on batteries. A system on a chip becomes a solution for designing and developing powerful mobile devices that consume lesser energy than traditional personal computers. SoCs are inherently more power-efficient because of their small sizes.
Maximizes the Benefits of RISC and ARM Architecture
Smartphones and tablet computers are now relatively as powerful as entry-level to mid-range level personal computers. They can accomplish the same tasks despite having SoCs that are considerably smaller than the hardware components of traditional computers. However, these mobile consumer electronic devices are built on different technologies.
Mobile devices are fundamentally on a separate category and territory than traditional personal computers. What makes a system on a chip an adequate component of smartphones and tablet computers is that they utilize and maximize the benefits of reduced instruction set computer or RISC and ARM architectures.
Most traditional computers, including Windows-based PCs and Intel-based macOS computers, are built on complex instruction set computer or CISC and x86 architectures. These architectures are powerful and complex, thereby requiring power-hungry processors. RISK and ARM architectures run smaller and simpler instruction sets to make processing easier.
Advances Further Consumer Electronics Technology
Developments in SoC technology have resulted in further developments in consumer electronics technology, thereby advancing further the entire consumer electronics industry. TO reiterate, another key advantage of SoCs is that they consist of most or all of the major components of a computer system within a small footprint.
The developments in system on a chip tech comes from several manufacturing improvements, including advanced process nodes such as the 5nm manufacturing process node. These improvements have allowed chipmakers to integrate more advanced components and technologies in a single chip while improving overall performance.
Note that mobile SoCs based on the 5nm process node such as the A14 Bionic, Snapdragon 888, and Kirin 9000 are equipped not only with CPUs and GPUs but also with Ai or ML engines, image signal processors, LTE and 5G network capabilities, including support for Sub-6 and mmWave 5G, and advanced Wi-Fi and Bluetooth radios, among others.
One of the hallmarks of SoC technology is the introduction of the Apple M1 chip in October 2020. Found in the 2020 MacBook Pro, 2020 MacBook Air, the 2020 Mac Mini, and the 2021 iPad Pro series, the M1 chip demonstrates the use of the ARM architecture in powerful computing devices, and the attempt of Apple to shift from x86 Intel architecture.
Cons of SoC: Disadvantages of System on a Chip
Challenges in Manufacturing a System on a Chip
There are several issues and challenges that manufacturers need to overcome when designing and mass-producing SoCs. For starters, they need to optimize waste heat output on the chips. The main problem with integrated circuits is that they are susceptible to unsustainable energy density levels, much of these would be dissipated in the form of heat.
One of the causes of overheating in electronic components and consumer electronic devices is the physical dimension. Smaller components packed in closer proximity to one another will not only accumulate high energy density levels but would also accumulate heat due to the reduced thermal flow. Factoring in heat management in the design of SoCs is essential.
Another manufacturing issue centers on improving chip performance without increasing its size and affecting power consumption. The chipmaking industry has successfully implemented the 7nm and 5nm manufacturing process nodes. However, this achievement has cost implications due to added manufacturing complications and material requirements.
Pushing the process node beyond 5nm and 3nm can be challenging because it would mark another attempt to stretch the limits of Moore’s Law. Note that designing and producing 7nm and 5nm chips have critical drawbacks. Even the chips themselves have inherent issues and limitations that have been addressed using novel but costly manufacturing solutions.
Individual SoC Components are not Upgradeable
Upgradeability is a critical disadvantage of a system on a chip. To be specific, while the entire chip is theoretically detachable and upgradeable, its individual components are not. This drawback comes from the fact that these components are integrated into a single system that makes removing each of them impossible.
To illustrate, suppose a user decides to upgrade the memory of his or her device, or perhaps its GPU, he or she has no choice but to purchase a new SoC. However, in most scenarios, such as in the case of almost all smartphones and tablet computers, the only option is for him or her to buy a new device. The same is true for the M1 Mac computers.
Same with upgradeability, the impossibility of removing individual hardware components lowers the overall reparability of the device. A damaged component would mean a damaged system. Even a small damage to a particular area can affect the overall integrity of the chip. Remember that the components within an SoC are interconnected.
A device based on a traditional computer architecture is not only upgradeable but has a higher reparability score. Users can readily purchase individual components to upgrade their devices whenever and however they see fit, and in consideration of compatibility. This means an individual can upgrade to a new graphics card without buying a new CPU.
Traditional Architecture Remains More Powerful
Generally speaking, and in consideration of current technologies, the fact remains that traditional computer architecture based on circuit boards is more powerful than systems based on SoCs. There are several reasons for this. The first is that motherboards in personal computers have separate components with very dedicated functionalities.
For example, graphic cards currently available in the market are more powerful than the integrated graphic processing units found in several CPUs, thereby making them ideal for high-graphics and resource-intensive use cases such as extensive video editing and computer animation generation, as well as high-end video gaming.
Traditional computer architecture is not only ideal but also practical for playing video games on a Windows-based PC. Gamers can opt to buy a desktop computer and have its components upgraded in the future to meet the evolving requirements of advanced game titles. Gaming consoles are also based on traditional computer architecture.
Remember that SoCs are impractical because they are not upgradable. For enthusiast-level users of personal computers, upgradability is essential, especially if they are into building and modifying their devices to meet current and future specific use case requirements and scenarios. A user can easily purchase a better RAM or a dedicated sound card if budget permits.
A Note on the Advantages of CISC and x86 Architectures
SoCs can be designed to run instructions based on complex instruction set computer or CISC and x86 architectures in theory. However, RISC and ARM architectures are still ideal because integrated chips are produced and deployed in consideration of simpler instruction processing, overall processing efficiency, and lower power consumption.
Mobile SoCs available in the market would not be able to run instructions based on CISC and x86 architectures. As an example, a Snapdragon 888 or an Apple M1 chip is unusable to operating systems and software based on the x86 architecture simply because they are designed to run on operating systems and software based on ARM.
CISC and x86 architectures have notable advantages over plain RISC and ARM. A CISC processor can perform multiple operations per single operation. It also requires an assembly language that is easy to implement. The hardware does more of the work in decoding instructions. The complexity of instructions means less memory requirement.
In a Nutshell: Advantages and Disadvantages of System on a Chip
An SoC has notable advantages to include smaller physical dimensions, more efficient power consumption, and integration of most or all critical computer system components, thus making it an ideal integral hardware for mobile devices. The remarkable performance of these chips also comes from the reduced processing latency owing to the close proximity of individual integrated components, as well as from better throughput due to the density of transistors in a given area. These characteristics have also advanced further the consumer electronics industry, as seen in the introduction of the desktop-class SoC Apple M1.
However, when compared to traditional computer system architecture based on motherboards and separate hardware components, it has notable disadvantages as well. First, SoCs are not upgradable because their components are not removable, thereby making them impractical for specific use case situations. High-performance computers are still based on central circuit boards because individual and separated components remain powerful than the integrated components of a particular chip. The manufacturing challenges of SoCs represent production drawbacks due to technical and cost implications.
FURTHER READINGS AND REFERENCES
- Asghar, M. N. 2020. “A Review of ARM Processor Architecture History, Progress and Applications.” Journal of Applied and Emerging Sciences. 10(2): 171. DOI: 36785/jaes.102446
- De Man, H. 1999. “System-on-Chip Design: Impact on Education and Research.” IEEE Design & Test of Computers. 16(3): 11-19. DOI: 1109/54.785820
- Ibrahim, D. 2015. “Microcomputer Systems.” In PIC32 Microcontrollers and the Digilent ChipKIT. Elsevier. DOI: 1016/b978-0-08-099934-0.00001-6
- Koo, K., Rho, G. S., Kwon, W. H., Park, J., and Chang, N. 1998. Architectural Design of a RISC Processor for Programmable Logic Controllers. Journal of Systems Architecture. 44(5): 311-325. DOI: 10.1016/s1383-7621(97)00011-8