Apple Silicon — The Move Away From Intel to ARM
One of the keynote presentations at Apple’s WWDC 2020 announced the shift from Intel processors to an ARM-based processor for the Mac computer lineup. The term Apple Silicon was applied to this new processor design, but in reality it can be applied to Apple’s current line of SoC (System-on-a-Chip)and SiP (System-in-Package) used in their mobile (e.g. iPhone) and tablet (e.g. iPad) devices. Apple had already used the ARM architecture in processors used for the iPhone, Apple Watch and even the Apple TV. Bringing it to the MacBook is going to be quite interesting.
Apple will design the processor for the new Macs, but the manufacturing may go to chip builders like TSMC. The Apple Newton was one of the earliest devices to use an ARM processor (ARM 610). The early ARM chips Apple designed were manufactured by Samsung for early revisions of the iPhone and iPod Touch. The “A” series chips were manufactured by TSMC, giving us fast processors like the A13 Bionic on the iPhone. The “S” series was specifically developed for the Apple Watch and contracted to Samsung. The performance of Apple’s ARM-based processors have been comparable to a standard desktop computer, and even surpassing that of low-budget entry-level brands. That makes plenty of sense to begin manufacturing chips for their own line of computers like the iMac and MacBook Pro.
PowerPC to x86
Oddly enough, prior to Intel the Apple computer architecture was based on the PowerPC RISC design, which is what the ARM architecture is based on. If Apple computers were originally using RISC, why did they shift to intel’s 64-bit x86 instruction set? It was due to decisions from leadership to move toward the PC mainstream, which was dominated by intel. With an x86 core CPU, the Mac and other lines of Apple computers can run popular applications without porting it to a different CPU microarchitecture.
Prior to the shift, Apple was also experiencing thermal issues with the PowerPC G5 processors. Apple’s engineers were having difficulty keeping energy consumption low for laptop and mobile devices, so that means more heat generated and less battery time. The designers at Apple needed an alternative processor that was meant for low energy consumption and thinner form factor devices. This led to the adoption of more efficient intel-based processors.
In 2006, the first generation intel-based Apple computers were released with Mac OS X 10.4.4 Tiger. To smoothen the transition, the PowerPC was still supported by their operating systems. Apple also created Rosetta, which allowed dynamic binary translation for applications built for the PowerPC to run on the intel-based processor. The iMac and MacBook Pro were the first line of Apple products to get an intel-based processor. With the release of Mac OS X 10.6 Snow Leopard, the PowerPC support was dropped meaning the OS can only run on intel-based processors.
The Move To ARM
Apple’s announcement of future generation Macintoshes running their own designed ARM processors may not have surprised many in the industry. There were already indications of this since Apple began developing processors for their mobiles and tablet products. It was easy to envision the jump to an ARM-based architecture for all their products because it brings about full vertical integration from design to the final product. In the past Apple had to wait for Intel when developing products in order to get the design specifications of the chips. That will all change once Apple has shifted to developing their own processors. Apple will have control of all aspects of hardware and software design, so their developers can optimize the product much further than what they could do with intel’s chips.
With ARM-based MacBooks and iMacs, apps for the iPhone and iPad can be run natively without major modifications when built in ARM64 code. This integrates the product line from the Apple Watch all the way to the high-end iMac Pro or Mac Pro workstation running on the same platform. At the moment the applications you have on the MacBook Pro cannot run on the iPhone. You would have to get the correct version of the application for the devices. The integration of a common architecture streamlines interoperability in the Apple ecosystem.
Like the purpose of the original Rosetta, Apple will allow existing applications to run on the ARM-based processors through emulation with Rosetta 2. This will be through a format called Universal 2, which will allow developers to build the application and compile it for different operating systems. Native apps will have the advantage of course, since it doesn’t need to undergo any translation which creates an overhead. To help with the move to Apple Silicon, a Developer Transition Kit has been made available to allow developers to begin updating current applications to run on the new ARM-based processors.
Advantages Of ARM
The advantages of using ARM-based processor architectures goes back to the benefits of RISC design methodology. When Apple originally moved to the intel-based processor, it was due to lower energy consumption which can improve battery time and thermal efficiency. The ARM architecture was designed for the lowest possible energy consumption while maintaining the highest processing power. This is more efficient so it allows longer lasting battery time for devices and longer battery life in general.
When it comes to multi-tasking applications, the RISC design would fall short to intel-based processors. This is because intel’s x86 instruction set was designed for general purpose computing. They can do many things at the same time with multi-threading support. For example, a laptop running an intel-based CPU can run a word processor, a browser with several tabs open and stream video without difficulty while a smartphone cannot do this. This requires more energy though, and that consumes more battery.
ARM is based on the RISC principle of performing operations using simple instruction sets, which require less circuitry. In intel-based processors that use the CISC principle, microcodes are used to execute instructions. This requires building a ROM circuit to hold these instructions as the chip uses a complex algorithm to schedule and execute them. RISC chips don’t use microcode. This allows engineers to develop smaller sized chips at lower costs and lower energy consumption as well.
While intel-based processors are powerful but energy hungry, the ARM-based processors are powerful and energy efficient. This is through implementing energy efficient cores to supplement performance cores in the CPU. Think of ARM as various components working together to process data in a more efficient manner, while intel-based processors process data in a more complex manner. This is also called the Power Per Watt (PPW) indicator, in which a more efficient processor is able to complete instruction sets in as little time as possible (e.g. instructions per clock cycle) by requiring less power at their peak performance.
To further show the efficiency of ARM circuits, looking at the dynamic power consumption of a CPU’s logic gates is an important factor in logic gate switching activity. This is the relation of Power ( P ) with Voltage ( V ), Capacitance ( C ) and Frequency ( f ):
From the equation, voltage reduction is clearly a way of reducing power consumption in circuits. With operating voltages as low as 1.2 V (< 5 W energy), this is a significant feature of low power consumption in ARM-based processors. ARM chips achieve efficiency due to smaller transistor sizes.
RISC and CISC may have a difference in design philosophy, but they have actually adopted features from each other. RISC developers have since added more instruction sets to their architecture, while CISC developers have tried to execute instruction pipelines like in a RISC based processor. This has made RISC design like ARM more capable of running on computer devices other than smartphones and tablets. How Apple will deliver that performance from ARM processors is what remains to be seen.
Synopsis
Apple’s move to ARM could affect the development of products by other manufacturers, since Apple is a leader in the industry. What Apple does, others tend to follow. In this case computer industry giants like Dell, Lenovo and HP are keeping an eye on what is happening. If they also shift toward an ARM architecture for their future product line, it is another paradigm shift in computing. That is likely not to happen immediately though. For legacy users, there is no need to worry. Apple will continue its support of existing intel-based products, as announced during their WWDC 2020 keynote.
Apple will be able to optimize performance and provide more battery time for its laptops like the MacBook and MacBook Pro. For the iMac and Mac Pro, it means that full vertical integration would allow Apple complete control of design and development to more improved performance for Apple software like Final Cut Pro. There is no more need to wait for Intel to come up with improvements and updates to the processors. This can also lead to lower priced Apple products, since the chips no longer need to be purchased from Intel. Apple also improves performance with the integration of a Neural Engine for advanced AI capabilities, cryptography acceleration and high-performance GPU.
Overall, this means a common architecture for Apple products. This will allow the iOS, WatchOS, iPadOS, tvOS and macOS to work seamlessly with each other, like one continuous interface regardless of the device being used. What runs on an iPad will run on an iPhone and MacBook Pro without needing a separate version of the application. This will provide a more pleasant user experience for those already on Apple’s ecosystem, making it more interoperable and convenient.
