Back on 22 June Apple revealed its plan to transition away from Intel processors inside the Mac to its own processors - a process that it expects to take up to two years, with the first Macs with Apple processors shipping by the end of 2020.
This isn't the first time Apple has undergone a move away from one brand of processor to another (back in June 2005 Apple announced the move from Power PC to Intel) but this is the first time Apple will be making the processors itself.
In this article we're going to attempt to determine how Apple's processors will compare to the current Intel processors, but before we do that there are a few questions we need to address.
First up we need to look at why Apple is ditching Intel as its processor supplier. Then we'll address some of the reasons why Apple wants to use its own chips - while explaining what the difference between Apple Silicon and ARM is for those of you who are thinking that this is a question of Intel versus ARM. We'll also consider how big a risk it is for Apple to move to its own silicon and why Apple is willing to take that risk.
Once we have run through all of that we should be able to make an informed decision about how Apple Silicon will compare to Intel - and whether you should wait for the first Apple Silicon Mac, or if it would be wise to wait for the next generation (read: Should I buy an Intel Mac for more information).
Why is Apple ditching Intel?
The move away from Intel has been rumoured for some time. As long ago as 2015 analyst Ming-Chi Kuo (back when he was at KGI) predicted that Apple would start to design its own processors for Macs within the next few years. That prediction was based on a theory that Apple's own processors (at that time the A9 and A10 chips that featured in iPhones and iPads) would soon be good enough for the Mac. It was suggested that moving to its own processors would give "Apple better control the timing of Mac launches and Mac product features."
Then in April 2018 a Bloomberg report titled "Apple Plans to Use Its Own Chips in Macs From 2020" revealed details of an initiative at Apple, codenamed Kalamata, that would see Apple "able to more tightly integrate new hardware and software, potentially resulting in systems with better battery life." The initiative was part of a strategy to make all of Apple's devices work more seamlessly together, according to "people familiar with Apple's plans".
Back in October 2018 Kuo (by then at TFI Securities) suggested that the move from Intel to a custom ARM based chip would offer a number of benefits to Apple, namely: not being dependent on Intel's processor release schedule, increased profit margins, and the potential for improved market share if Apple passed its savings onto the consumer in the form of reduced Mac prices.
On the subject of not being so dependent on Intel: Intel has struggled over the past few years as it attempted to reduce transistor size to 10nm (with a view to fitting more transistors on a processor). The subsequent delays, and the cancellation of the Cannon Lake processor generation, no doubt frustrated Apple. But that was only part of the problem with Intel. Former Intel engineer François Piednoël told PC Gamer that Apple was becoming irritated by bugs: "Apple became the number one filer of problems in the architecture".
Apple has outlined its own reasons for the move away from Intel. Apple highlighted the fact that the transition would "establish a common architecture across all Apple products, making it far easier for developers to write and optimise their apps for the entire ecosystem." Apple also mentioned the "advanced capabilities of Apple silicon" and its "industry-leading performance and powerful new technologies."
Other benefits highlighted by Apple include that developers will be able to "make their iOS and iPadOS apps available on the Mac without any modifications". The transition will also "give the Mac industry-leading performance per watt and higher performance GPUs - enabling app developers to write even more powerful pro apps and high-end games."
Apple also explains that, thanks to the move, developers will gain access to "technologies such as the Neural Engine" which will "make the Mac an amazing platform for developers to use machine learning."
There's one other reason why now is the time to move away from Intel and it actually links back to the reason why Apple moved from Power PC to Intel back in 2005-6. Back then Intel was the dominant force, supported by the popularity of Windows bringing investment that PowerPC couldn't compete with. Fifteen years later and PC sales have plateaued and all the growth is in mobile devices. And the mobile devices seeing the most growth just happen to have chips designed by Apple.
We have more details about Apple's switch to its own processors here: Complete Guide To Apple Silicon And Apple's ARM Mac plans.
Isn't Apple moving to ARM?
You may have heard the rumours that Apple was ditching Intel long before Apple announced the transition to Apple Silicon, in which case you might be thinking isn't Apple using ARM processors in its new Macs?
Apple has made it clear that this is a move to Apple Silicon, not a move to ARM. Apple has developed its own System on Chips (SoCs) based on ARM, but describing the resulting processors as ARM would be like saying that macOS is Unix due to the fact that it is based on Unix. There is a lot more to Apple's series of chips than ARM, and that's probably why Apple's SoCs have proven to be superior to other ARM-based processors (such as the Snapdragon suite from Qualcomm).
Apple's SoCs offer advanced features that aren't available on other ARM processors, such as advanced power management, machine learning, the Secure Enclave, the Neural Engine, Apple's own GPU, and much more.
That's not to say that ARM brings nothing to the party. ARM is based on RISC (Reduced Instruction Set Computing) while Intel (also known as x86) is based on CISC (Complex Instruction Set Computing). Remember we mentioned transistors earlier as one of the challenges Intel was facing in trying to reduce their size? Processors based on the RISC architecture don't require as many transistors as CISC based processors - this lowers cost, power consumption and means that less heat is produced.
These benefits mean that ARM is well suited to small and light battery-powered devices such as smartphones and tablets, but also laptops and even servers. In fact there is a ARM based super computer - the world's most powerful supercomputer, Fugaku, runs on a Fujitsu-designed Arm SoC.
The existence of this super computer does suggest that fears that switching to ARM processors means that Apple won't be able to produce adequate workstation processors for the Mac Pro.
Having spent the past few years attempting to win pro creative back to the platform it would be a big surprise if Apple was to now leave this group out in the cold. There should be no reason why Apple can't use it's own expertise to create ARM-based Apple processors that will meet the needs of these users.
However, there is another device also aimed at creative pros that has been at a huge disadvantage due to its ARM processor: Microsoft's Surface Pro X. While the Surface does at least run Windows (getting Windows running on ARM was a challenge in itself) there is a shortage of apps that run on the ARM-based processor. Microsoft provided an emulator that would convert 32-bit Windows apps to run on the Surface Pro, but it can't convert modern 64-bit apps meaning that apps such as Adobe Lightroom can't even run on the Surface Pro X (reviewed here).
Luckily Apple has had the foresight to make sure that doesn't happen. We already know that Adobe is committed to getting its apps up and running on Apple Silicon - Apple actually showed Adobe Photoshop and Lightroom running natively on Apple Silicon during its WWDC keynote presentation. Plus Apple has its own solution to make sure that all applications will work on day one: Rosetta 2.
Anyone who is concerned that ARM processors aren't suitable for professional Macs should take note that the processors Apple is developing should not be disadvantaged in the way that the ARM processors that have already been seen on Windows laptops have been. Hopefully we can trust Apple not to let the pro creative market down again.
What about graphics?
The other concern creative pros have is the graphics capabilities of these new Macs. Right now some Macs ship with Intel's integrated graphics and others offer discrete AMD graphics, such as the AMD Radeon Pro 5600M GPU that is an option for the 16in MacBook Pro.
When Silicon Macs arrive Apple says that they will support Apple GPUs. While it's probable that Apple's integrated graphics will offer a boost compared to the Intel integrated graphics, there are concerns that Apple intends to replace AMD GPUs with its own.
It shouldn't be a surprise that Apple can design a GPU. Apple already uses its own GPUs inside iPads and iPhones with a lot of success, and Apple is setting expectations high - claiming that we can expect "higher performance GPUs" inside the new Apple Silicon Macs. The company said that the transition will lead to more powerful games on the Mac - which is quite a claim since the Mac has long been considered inadequate for gaming.
We've already seen evidence of this: During the WWDC keynote Apple demonstrated Shadow of the Tomb Raider running smoothly under Rosetta 2 on a special edition Mac mini featuring the A12Z usually found in the iPad Pro. Unity has also committed to working with Apple to get its Standalone player up and running on Apple silicon-powered Macs.
It's not just games: Maxon has also committed to getting Cinema 4D running on Apple Silicon and Apple demonstrated Cinema 4D running via Rosetta 2 during the keynote.
This probably won't stop concerns that integrated Apple GPUs can't possibly complete with the discrete GPUS of AMD and Nvidia. That's probably why Apple said the following in this developer document: "Don't assume a discrete GPU means better performance," and "The integrated GPU in Apple processors is optimized for high performance graphics tasks".
For more clarity on this, during a developer session at WWDC Apple discussed the GPU architecture of its new Macs. Apple uses Tile Based Deferred Rendering (TBDR) rather than the Immediate Mode Rendering (IMR) as used by Intel, Nvidia and AMD GPUs. TBDR captures the entire scene, splits it into small regions, processes them separately and therefore doesn't require as much memory bandwidth. The scene isn't be rendered until occluded pixels (pixels that shouldn't be visible) are rejected. In IMR the whole scene is rendered first, prior to the rejecting the unnecessary pixels. That latter process requires a lot more bandwidth.
If you are wondering how the next round of Macs with Apple Silcon might shape up read about them here:
Pros and cons of Apple Silicon vs Intel
We've already covered many of the advantages of moving to Apple Silicon from Intel and some of the perceived disadvantages (specifically the experience to date of Windows on ARM). We'll run though the pros and cons of the transition below though.
- Apple will no longer be frustrated by Intel's bugs and delays (such as the cancellation of the Cannon Lake processor generation).
- Apple's iPhone & iPad processors are already good enough for the Mac.
- Apple will have more control over timing of Mac launches.
- Apple won't be tied to Intel's processor release schedule.
- Apple will be able to tightly integrate hardware and software.
- Better battery life should be possible.
- All of Apple's devices will work seamlessly together.
- Apple should see improved profit margins and might be able to pass savings onto the consumer in the form of reduced Mac prices (although it probably won't).
- It will be easier for developers to optimise apps for the whole ecosystem: Mac, iPhone and iPad
- Industry-leading performance per watt.
- Higher performance GPUs should mean more powerful pro apps and high-end games.
- Access to the Neural Engine should allow developers to use machine learning.
- Apple's leading the industry with its chips for smartphones and tablets and can do the same for the Mac.
- Unlike other ARM-based chips, Apple's processors will have access to features such as advanced power management, machine learning, the Secure Enclave, the Neural Engine, Apple's own GPU, and much more.
- ARM-based processors don't require as many transistors which reduces power consumption and means that less heat is produced.
- The world's most powerful supercomputer, Fugaku, runs on the A64FX, a Fujitsu-designed Arm SoC.
- Apple has been making processors since it designed the A4 chip in 2010, so it has 10 years of experience. You could argue that Intel has many more years under its belt.
- The first generation of any product has risks associated with it. Apple Silicon is new and untested on the Mac. We don't know what Apple will be able to achieve.
- The ARM-based Surface Pro has a number of issues due to incompatible software - especially consumer pro apps.
- ARM isn't compatible with x86 or x64 software, which means that some Windows apps won't run without an emulator, but this is unlikely to be a significant issue for Mac users.
- One of the benefits of Apple's move to Intel back in 2006 was that it meant that Mac users could run Windows on the Mac. This opened the Mac up to people who were hesitant about leaving Windows behind.
- While ARM (and AMD) are closing in on it, Intel is still dominant and this means that developers will continue to design apps for Intel.
It looks like there are many advantages to moving from Intel to its own silicon, and with Apple's processors already rivalling Intel processors this is the time to do it. With Apple making the silicon we should no longer be waiting for years for Apple to update Macs with the latest processors. But this is about more than the processors: the benefits of having a unified architecture will bring iOS apps to the Mac, which could in itself change the way we use our Macs. This could be the beginning of the next transformation in computing.