WebProNews
WWDC 2020 Part 4: Apple’s Custom Silicon—How It Works and What It Means
As we previously reported, Tim Cook announced what many had predicted: The Mac is officially moving to Apple’s custom silicon.
Apple has been rumored to be moving working on moving the Mac to ARM processors for years, especially as Intel has struggled to keep up with industry developments. In many ways, Apple’s current situation mirrors the situation it found itself in with its last line of chips, the PowerPC semiconductors.
Apple’s Semiconductor History
Used by the AIM alliance (Apple, IBM and Motorola), PowerPC was a RISC-based architecture that had a number of advantages over Intel and other x86 lines. Ultimately, however, Motorola and IBM fell behind Intel, in terms of development and processor speed. Regardless of how much more advanced the PowerPC chips may have been, Intel’s sheer processing speed eventually surpassed it.
To make matters worse, IBM was never able to solve the issues with heat. This was especially evident with the PowerPC G5 line of chips. While extremely powerful for desktop computing, it could never be used in a laptop, consuming too much power and producing too much heat.
After several years of falling behind, at the mercy of IBM and Motorola, Apple jumped ship to Intel. Once again, however, Apple is in the same boat. Intel has been struggling to keep up with demand and move to 10nm processors, leaving room for its old rival, AMD, to make significant headway.
To make matters even worse, Apple has had issues with some recent MacBook Pro models not being able to sustain high-speed operations because of the heat generated by the Intel processors. Instead, the machines have had to throttle their performance, in some cases making top-end models run slower than low-end and mid-level MacBook Pros.
ARM Chips To the Rescue
The game-changing element for Apple is the rise of ARM chips, which the company uses in its iPhones and iPads. Arm Holding was a joint venture between Apple and Acorn.
Unlike Intel or AMD, Arm doesn’t manufacture its own chips. Instead, it designs and licenses them for its customers to manufacture on their own. The company offers different license level, some of which allow customers to modify the designs to better suit their needs.
Apple, however, has the broadest license of all, essentially allowing it to do whatever it wants with Arm’s designs. The results are industry-leading chips that provide unrivaled performance. Case in point is Apple’s recent iPhone SE. Despite having an under-clocked A13 Bionic chip, it significantly outperforms flagship Android phones.
What’s more, Apple’s manufacturing process has been able to keep up with the demand for iPhones and iPads, which far exceed its Mac base. As a result, the writing has been on the wall for some time that Apple would eventually switch to its own custom silicon, taking its fate back into its own hands once again.
Performance and Power
One of the biggest benefits of moving to its own silicon is the ability to deliver Macs that offer better performance while consuming less power. For example, the current generation iPad Pro delivers better performance than most PC laptops on the market, and does so without a single fan to help manage heat.
In the large case of a MacBook, iMac or Mac Pro, the performance possibilities are thrilling. During the demo, Craig Federighi showed one of the new Macs running Final Cut Pro. Not only could he edit and add effects while the video was playing, but Final Cut could run three simultaneous 4K streams.
Developers and Compatibility
While any processor change is a monumental undertaking, Apple has a history of pulling it off, with this being the third such transition.
Federighi highlighted the work the company has done to help ease the process for developers, with many of them able to update their apps with just a few days of work. Microsoft and Adobe have already ported their software to run on the new architecture.
In addition, the company will make it possible to ship Universal apps that contain binaries for both Intel and Apple’s CPUs. Apple has labeled this Universal 2, and improved version of the Universal apps that contained PowerPC and Intel binaries during that migration.
In addition, macOS will include Rosetta 2, a translation layer that will run Intel-based software that hasn’t been updated yet. Again, Apple first introduced Rosetta during the Intel migration, allowing the Intel-based Macs to run older PowerPC-based software.
The new version of Rosetta is a significant upgrade, however. Whereas the original Rosetta translated an app at runtime, Rosetta 2 will translate the time of install. This will significantly improve performance of these older apps. Meantime, a just-in-time (JIT) compiler is still available if needed, such as when working with plugins.
In addition, Andreas Wendker demoed Parallels running Linux. This is an important factor, as it shows that virtualization is alive and well on Apple’s silicon.
He also demoed Shadow of the Tomb Raider running in Rosetta translation mode. Despite running at 1080p, the game was flawlessly smooth.
As an added bonus, the new ARM-based Macs will be able to run iOS and iPadOS without modification, thanks to the shared architecture.
Custom Silicon—A new Era of Mac Computing
Without a doubt, Apple’s move to its own custom silicon has been a long time coming. The move will help usher in a new era of Mac computing, with Apple firmly in charge of its own processor development.
As a company that has built its experiencing on tight control of the hardware and software, this is a move that will help usher in a new wave of Mac innovation.
WWDC 2020 Part 4: Apple’s Custom Silicon—How It Works and What It Means
Matt Milano
from WebProNews https://ift.tt/2CwyIy9
No comments:
Post a Comment