Apple’s phasing out Intel processors in Macs in favor of Apple silicon. To understand what that means for Mac performance, AppleInsider has looked back at past and current A-series chips and compared them to Intel CPUs.
Mike Peterson for AppleInsider:
The A10X Fusion is in the same ballpark as a 13-inch MacBook Pro model from 2017. The specific variant we’re comparing is the mid-range configuration with an Intel Core i5 processor, which retailed for $1,499. The aforementioned 13-inch MacBook Pro clocks in with a single-core Geekbench 5 score of 850 and a multi-core score of 1972, meaning it’s actually slightly slower in multi-core performance despite being more expensive than the iPad Pro…
The A11 Bionic came in with a 917 single-core and 2350 multi-core score in Geekbench 5 benchmark testing. While that device retailed for $999, the A11 Bionic could be found for cheaper in the iPhone 8 series. The entry-level 2020 MacBook Air, equipped with an Intel Core i3-1000NG4, benchmarks similarly…
The A12 Bionic [is] on roughly the same footing at the 2017 21.5-inch iMac equipped with a 3GHz Intel Core i5-7400 processor. That device started at $1,099.
The A13 Bionic averages a single-core score of 1325 and a multi-core score of 3382… If you want similar performance in a Mac, you’ll probably want to take a look at the 13-inch MacBook Pro with an 8th-generation Intel Core i5-8257U processor… On Geekbench 5, it comes in lower than the iPhone in single-core scores but slightly higher in multi-core with 1012 and 3676, respectively.
The A12X and A12Z Bionic both benchmark around 1115 in single-core testing, but they clock in as the fastest iOS-based devices in multi-core benchmarks with a high score of 4626. On an Apple Developer Transition Kit running Geekbench 5 natively, the scores are roughly similar with single-core scores around 1005 and multi-core scores around 4555… If you want similar single- and multi-core performance in an Intel-based chip, the way to get it is the mid-range 16-inch MacBook Pro with a 10th-generation Intel Core i7 processor. It retails for $2,399.
MacDailyNews Take: There is much more in the full article. Please read it here.
We haven’t seen anything yet! Apple’s 5nm A14 is going to shame Intel to its, uh, core.
Buh-bye, Intel slug! Intel served its purpose, but has been a boat anchor for years. Hello, Apple-designed ARM-based Macs! — MacDailyNews, April 23, 2020
But how does it do against the New Mac Pro?!?!? Apple finally realeases a proper Mac Pro and then they cut it off at the knees by changing the CPU lesss than 6 months later.
Avid still hasn’t gotten QuickTime to work properly in Catalina yet?!?
First, while the benchmarks quoted in the article are somewhat irrelevant because a full blown A-series chip for a Mac will require lots and lots of additional capabilities that will inherently slow the chips down a bit and require more power (and heat dissipation capability), no one is saying that the Mac Pro is switching to an A-series chip any time soon. It is very likely that the Mac Pro will be among the last, if not THE last Mac to make the transition.
Further, Apple is known for supporting systems for several years. In all likelihood there will be one more Intel based Mac Pro coming out some time in 2021. Apple likely will support that box for five or more years. So even people buying an Intel based Mac Pro in the second half of 2021 very likely will not have to be too worried for multiple years after that.
And, who knows what the state of desktop computing, let alone the Mac Pro, will be in come 2024 or 2025? Not even the best prognosticators at Apple or in the Media really have any idea.
Finally, Avid is not responsible for QuickTime, Apple is. Issues with Avid working with Apple software is a combination of Apple’s and Avid’s problems, not Avid’s alone.
Avid still around? The Quark of video editing industry.
You’re putting the entire blame on Apple, but there is a good possibility Intel wasn’t able to deliver on smaller node Xeon processors as promised. I’m just saying if we don’t know exactly what’s going on behind the scenes, it difficult to put the blame in the right place. It is now very obvious Intel is messing up its processor timeline. Many corporations get contracts based on future products and that can really burn them if those timelines are pushed back.
I would hardly think buying an Intel Mac Pro would be a total loss even if a newer model came out using Apple Silicon. However, we would have to see exactly what an Apple Silicon Mac Pro could do in terms of processing power.
These Apple Silicon chips are low power. They can throw four of these things into a Mac without breaking a sweat. One just for thunderbolt in/out.
My question has always been if the A-series SoCs were designed to support multi-processor operation (each SoC processors containing multiple cores, or course). If so, then rough equality of benchmarks between an A12X/A12Z and an Intel Core i5 is a great starting point because Apple could then throw two or more A12X/A12Z SoCs in a Mac to boost performance, like with the G5 and G4 processors, (but not the G3). Given the thermal efficiency of the A-series SoCs, multi-processor configurations seem like the way to go.
If the current A-series SoC design does not support multi-processor configurations, then Apple will only be able to deploy A-series SoCs on lower-end Macs until it develops faster versions (A13X, A14X) and/or a multi-processor compatible SoC design.
Apple is reportedly planning to roll out A-series processors across its Mac lineup over the next two years. If so, then you can bet that the higher end Macs will be using A13, A14, and A15 SoC variants beefed up similarly to the iPad Pro processors with extra GPU cores. Apple may even be working on some special Mac edition SoCs for the high end – hopefully multi-processor capable. I would like to see 32, 64, or even 128 A-series SoCs running in parallel on a Mac Pro Workstation beast to utterly annihilate the competition and destroy the benchmarks.
If anybody here followed the Apple transition from PowerPC to Intel, they would be familiar with the issues about transitioning. While straight benchmarks may be compelling, they do not reflect real world expectations. Apple will not have the new ARM-based Big Sur OS fully optimized for ARM. And judging by the last transition it will take several versions (years) for all the subsystems (and new replacement layers) to be fully native on the new platform. And a lot of the main productivity apps that form the core of many professional Mac users — Microsoft Office and Adobe Creative Cloud — will need to be mostly rewritten to take advantage of the new hardware and OS and run native. Many apps will be run in compatibility environments until obsolete. Or they will make use of parts of the OS that are yet to be native and suffer.
So, while the A series processors may have a lot of speed and thermal benefits, those will initially be offset by lack of full compatibility. The move from PowerPC to Intel had the same issues. And the first wave of Intel processors were similarly lauded by their power and thermal capabilities. But those weren’t obvious and fully exploited till several years down the road.
And finally, isn’t anybody else worried about the future of ARM’s ownership? I find it extremely risky that Apple is making this move without the future of ARM being fully transparent. I mean, does anybody think nVidia owning ARM would be a good thing? The two companies have been at each others throats for years. https://www.pcgamesn.com/nvidia/apple-cuda-toolkit-support
And how does Apple still claim “ownership of the technology” behind ARM, when it is not an owner, but just a licensee? Does owning the manufacturing of licensed technology represent that much of a benefit of just purchasing manufactured parts from another business that owns the patents?
Yes, Apple should engage in a consortium to own ARM. That they haven’t signaled their strategy to do so is troubling.
Couch money for Apple to buy ARM. Read a good article about Softbank selling ARM and why Apple might not want to buy it recently. But, all agreements have an endpoint.
As far as conversion, Apple has used ARM chips since the first iPhone and A-series chips since June 2010-thats 10 years folks. Apple thinks far ahead and probably has had a team on this for many, many years. I hope this conversion will be much more advanced and integrated than the previous 2.
Of course, I personally don’t need anything special other than chest-thumping, told-you-so bragging rights. lol
One difference between PPC-to-Intel and Intel-to-ARM is that the iPads and iPhones have been using those chips for years. So why port your intel code when you can just beef up your iPad code with new features?
The dark side of this is that we may loose a lot of things that iPads don’t do, like having a proper finder.
I’m guessing the CI/CD process for building macOS/iOS is much more refined than what it was during the PowerPC->Intel transition. It might be very possible they can “click a button” and after a few days they get a full blown OS image.