Apple has never been afraid to make radical changes. Whether it’s dumping the headphone jack on iPhones or getting rid of optical drives on Macs. They’ve also changed their fundamental CPU technology several times, with the shift to Intel merely being the latest one. It’s a bold move to break native code-compatibility with your entire existing software ecosystem, but Apple have done it when they’ve felt it was warranted. Now Apple is shifting once again, from Intel to their own custom ARM-based processors. With the first example of this “Apple silicon” Mac being the M1 chip.
Why The Change?
We’ll never know the exact insider reasons Apple chose to switch from Intel to their own CPUs, but it’s not a major mystery if you piece together all the factors we do know. The primary driver behind this change is Intel. Rather, Intel’s consistent failure to keep up with advancements in processor technology.
This has been a particularly bad year for Intel. They still can’t seem to push all the way to 10nm CPUs and have been forced to refine and rework their 14nm process over and over again. In the meantime, AMD’s latest Ryzen processors make use of 7nm technology. These processors are beating their Intel equivalents both in single-core instructions per clock and in multi-core performance. AMD is now the clear leader, with Intel’s main advantage being the ability to supply enough CPUs to meet demand, not provide the best CPUs.
This is a problem for Apple. For their Macbooks in particular, power-hungry and hot CPUs have been a major issue. From thermal throttling to poor battery life, Intel was doing Apple’s reputation no favors.
In the meantime, Apple’s ARM CPUs, especially those in the iPad Pro, stand toe-to-toe with desktop class CPUs found in mainstream laptops. That’s while running from battery power and without any form of active cooling. Increasing the size of these CPUs and giving them access to active cooling meant they could meet and exceed the performance of the Intel chips while improving battery life and thermals.
It’s also probably not a coincidence that this means more profit for Apple and much tighter control of their ecosystem.
Why The M1 is Special
The M1 Chip is based on the A-series SoCs found in the iPhones, iPads and other mobile Apple devices. It features the same CPU fundamental CPU cores, GPU cores and Neural Engine technology.
The M1 has eight cores, but it uses an asymmetrical design. Four of the cores are power efficient and work for basic tasks such as web-browsing or watching videos. The other four cores are high-performance cores that kick in for heavy tasks such as rendering, video encoding and gaming.
The M1 is also a highly-integrated system-on-a-chip. It incorporates the system RAM and cache as well. This means that the CPU and GPU have equal access to those resources. It also saves on interconnects and makes the entire system more power efficient.
Apple have added their own secret-sauce optimization and customization to the ARM design and the benchmarks seem pretty conclusive. In benchmarks such as Geekbench and Cinebench the M1 is faster in both single- and multi- core operations. When compared to Core i7 chips in other ultrabooks. It’s also posting better scores than the Core i9 in the MacBook Pro 16”. Although of course there are thermal constraints at play there.
The Burden of Rosetta 2
While the M1 seems to be a clear step up over Intel Macs in every category, it’s not a change without penalty. When Apple changed from PowerPC chips to Intel in the mid-2000s it was necessary to create a translation layer, so that old-generation apps would still work.
That software was called Rosetta. Although it did allow PowerPC code to run on x86 processors, it also came with a severe performance penalty.
Rosetta 2 is the new solution to allowing Intel Mac apps to run on these new M1 computers. The performance penalty is much less severe, but it varies from one app to the next. For general purpose apps that aren’t performance sensitive anyway, it makes little difference. However, for professional applications that need every bit of CPU power that penalty can be significant.
On the other hand, some applications are running just as quickly as they do on Intel Macs, through Rosetta. This is because the performance increase is such that it offsets the overhead of Rosetta.
That being said, it’s best that you wait for ARM native versions of professional applications, to get the best overall boost.
The Shift Is Inevitable
While it’s not the right time for everyone to make the change to an M1 system, it’s undoubtedly going to become the only choice moving ahead. What’s really exciting is the prospect of a massive, actively-cooled multi-core Apple chip powering the desktop Mac Pro of the future. If you need to buy a new Mac today, the M1 has few disadvantages apart from the Rosetta overhead. If you aren’t in dire need of a new Mac, it’s OK to hang onto that Intel system until your core applications have been ported to ARM.
