ARM MacBook Pro I don’t want an Overpriced Netbook.

[Maximara]

Something being available for Windows on ARM is irrelevant. For one thing, Apple Silicon is significantly more performant than ARM Cortex SOCs. For another, Apple demonstrated full Office running natively on Apple Silicon at WWDC - and they did it on multiple programs and did it more than once.

This whole discussion has entered FUD territory.

As I said before, unless you repurpose FUD to mean Fictional Unsupported Delusions things passed the standard Fear Uncertainty and Doubt meaning a long time ago.

 

[Joelist]

As I said before, unless you repurpose FUD to mean Fictional Unsupported Delusions things passed the standard Fear Uncertainty and Doubt meaning a long time ago.

I was going to use a different acronym like (FAR - FICTIONAL ALTERNATE REALITY) but I like your take on FUD.

 

[Yebubbleman]

Will a first Gen ARM based MacBook Pro Surpasss and Outperform the previous Intel based Version ? I’m gonna say No.

It May have Longer battery life, it will have WiFi 6 (intentionally left out of current gen) It May be less expensively But will it be “Faster”

Heck they can go from LPDDR4 to LPDDR5 RAM today by switching to ARM.

i feel like the link below we are getting clowned by the Swicth ti ARM, which won’t have critical support for Windows Boot Camp installations for Gaming or support for an eGPU.

You're not going to see the release of any Apple Silicon Mac that doesn't outperform its Intel predecessor equivalent. Apple knows that there are zillions of skeptics like you out there that they need to convince.

Furthermore, Apple Silicon SoCs from 2018 (A12X) already outperform all Mac minis, 21.5" iMacs, MacBook Airs, and 13" MacBook Pros that have ever existed. They are on course to outperform the remaining Macs by the end of their transition time given how fast they've advanced their SoCs.

So, yeah, it WILL be faster.

Yes, the loss of Boot Camp and x86 virtualization is huge and a massive bummer. Playing devil's advocate here, those things were never at the core of the Mac experience; they were just perks of using x86. Plus, there's no guarantee that we won't see Windows 10 for ARM64 make its way to Apple Silicon Macs in one form or another (be it native dual-booting or virtualization). And it's sounding like Microsoft may also be starting to work on adding x86-64 application support into Windows 10 for ARM64. I won't believe it until I see it, but certainly Microsoft needs to do SOMETHING to evangelize Windows 10 for ARM64. Otherwise, it won't have much of a point for existing.

 

[Nicole1980]

Try to have a little faith in the company, or at the very least, its economic modeling, to provide value to its clients. The switch to ARM will bring silicon investment into the arms of Apple, and this will help lead them toward being able to allocate more resources to R&D and the end product rather than being beholden to obnoxious lawsuits and hardware limitations. Never forget these are complex machines and a feat of engineering.

The removal of ports was a design decision with trade-offs. It does not save them tremendous money to exclude another port. A portless machine ought to be the goal in my eyes, and it has certainly been the prediction since the hour that carrying around your case of CDs was no longer that cool. Lol.

A vison of what the future 'should' be in some peoples eyes don't often really make sense in the real world.

I'm sure the people who watched Metropolis in the 1930's thought it absolutely made sense that eventually people would commute around town in propeller planes. Just like how you feel a future without ports makes perfect sense.

I'm not saying you're necessarily wrong, I'm just saying you need to question whether you're following what makes true sense in the real world versus how the current trajectory of things seems to be guiding your train of thought.

Make sense?

 

[Nicole1980]

You're not going to see the release of any Apple Silicon Mac that doesn't outperform its Intel predecessor equivalent. Apple knows that there are zillions of skeptics like you out there that they need to convince.

Furthermore, Apple Silicon SoCs from 2018 (A12X) already outperform all Mac minis, 21.5" iMacs, MacBook Airs, and 13" MacBook Pros that have ever existed. They are on course to outperform the remaining Macs by the end of their transition time given how fast they've advanced their SoCs.

So, yeah, it WILL be faster.

As a video editor, this is where things get murky.

When most people (and probably you) talk about 'performance' you're probably talking about Geekbench scores (or something similar). But.... Geekbench scores are heavily influenced by tricks like hyper threading and turbo boosting.
Well, when you have any kind of long heavy duty process that could really benefit from that 'speed' ... guess what - the turbo boost benefit that shows up in Geekbench disappears because things like turbo boost are only meant to run for short periods of time (ESPECIALLY on the typically 'thin' apple products), and on processes that only require a core or two. But no processor can stably deliver long term performance at anywhere near peak turbo boost speeds.

So, all this bravado talk about how well Apple silicon will perform? That to me comes with that huge question mark: Apple silicon may beat Intel in a 2 minute benchmark, but how will it compare when I'm rendering a 1.5 hour video and the heat the pressure on the processors drop them back to their 'stable' performance levels?

To me, that is where the real proof in the pudding lies, and certainly nothing I've seen has given any sense of what the real answer will be.

So don't show me another stupid Geekbench score. Show me what sustained - long term - performance looks like when comparing the two processors. Until I see that, all you're talk about Apple silicon trouncing Intel chips just feels like vacuous slight of hand.

 

[EdT]

You’re right, but I think we’re talking about slightly different things. I am referring specifically to the silicon. Not the software or it’s implementation necessarily, but how the silicon compares to intel.

Sure we can guess how the software and real-world implementation will fair, but that’s not the point of my post. My point is that when apple compares the newest model to the older model, especially when it comes to their silicon chips, they love to say Things like “20% faster than before”. Does that always translate to real-world performance? No. But it does show that overall the chip is still better than the year before and that’s apples entire MO with their custom silicon.

Other than the Mac mini (which was ruled by intels chips and their thermals) Apple always sells the newest models as being better than the previous years.

On a related note - a lot of people are skeptical about how powerful these chips are because they just announced the transition. People forget that Apple easily could have been working on this transition for over half a decade or more in their R&D labs. Just because it’s new to us does not mean Apple hasn’t tested hundreds of prototypes over years and years. Steve even said every version of OSX had an intel version too even during the PPC days - so just because Big Sur is the first official OS to work on ASi it doesn’t mean they haven’t had an ARM version of Mojave or high Sierra even.

Apple has done a lot of transitions. They are pretty good at them. And now that they have 100% control over the entire transition from top to bottom, a lot of the concerns mentioned may have already been ironed out during the Mojave or Catalina days.

But to get back to my point, I think Apple silicon chips are going to be at least as good as intel. BUT I don’t think Apple would have made such bold claims at WWDC if they were expecting to surpass intel chips (they already have in single core performance in the iPhone chip).

In the short term, I think I am talking about what a Mac can and can't do using x86 versus what its limits are with Apple Silicon. The potential growth of either system I couldn't say, but with Intel/AMD x86 the potential for massive improvement isn't there, and probably is for an ARM chip. But right now software is more advanced and functional for the x86. And I don't know what limits an ARM has or doesn't have in comparison or how long until the software catches up. Thats not just a hardware situation its a who is developing what and how hard is it to write ARM code to do what x86 code can already do.

 

[Joelist]

There is no such thing as "ARM code" and in fact there is no such thing as "Apple Silicon" code that is written by a developer (and again calling Apple Silicon ARM is not accurate). The basic difference between the current code and coding for ASi is mostly a recompile with a few cases of different security and/or addressing of some hardware.

 

[johngwheeler]

As a video editor, this is where things get murky.

When most people (and probably you) talk about 'performance' you're probably talking about Geekbench scores (or something similar). But.... Geekbench scores are heavily influenced by tricks like hyper threading and turbo boosting.
Well, when you have any kind of long heavy duty process that could really benefit from that 'speed' ... guess what - the turbo boost benefit that shows up in Geekbench disappears because things like turbo boost are only meant to run for short periods of time (ESPECIALLY on the typically 'thin' apple products), and on processes that only require a core or two. But no processor can stably deliver long term performance at anywhere near peak turbo boost speeds.

So, all this bravado talk about how well Apple silicon will perform? That to me comes with that huge question mark: Apple silicon may beat Intel in a 2 minute benchmark, but how will it compare when I'm rendering a 1.5 hour video and the heat the pressure on the processors drop them back to their 'stable' performance levels?

To me, that is where the real proof in the pudding lies, and certainly nothing I've seen has given any sense of what the real answer will be.

So don't show me another stupid Geekbench score. Show me what sustained - long term - performance looks like when comparing the two processors. Until I see that, all you're talk about Apple silicon trouncing Intel chips just feels like vacuous slight of hand.

Absolutely! The only meaningful benchmark for any given individual is whether you can do your specific tasks more quickly or more efficiently. This includes not just the time to complete a task, but the overall experience of working with the machine. Does the battery last longer? Is the user interface smoother or the file system faster? Does it run the software you need?

At present we have very little idea of how well ASi Macs will run MacOS and specific applications. Geekbench results based on existing iPad SoCs only provide a vague idea, and as you say, they may be skewed by the tests themselves.

Key metrics for me will be:
1) Is general MacOS responsiveness at least as good as my current MBP16
2) Do Safari & Chrome web browsers work well and are my plug-ins supported?
3) Does it run MS Office smoothly?
3b) Are other common productivity apps supported, e.g. Slack, Zoom, Skype, Teams
4) Can it play back and scrub 4K UHD H.265 or H.264 video at 60fps without dropped frames?
4b) Does it natively run Davinci Resolve & at least some non-Adobe photo editing apps?
5) Will it render videos (or perform ffmpeg encodes) faster than current equivalent-model Intel Macs?
6) Will it run common development software and frameworks such as Python, Node.JS, Java, MySQL, Docker containers?
7) Will it read/write NTFS formatted disks either natively of with 3rd party tools?

That covers my main needs. I'm sure there will be a few "gotchas" where a specific app doesn't run under Rosetta, so it will require some research to see if my tools are supported, and if not, whether the vendor plans to support them on Apple Silicon.

 

[johngwheeler]

There is no such thing as "ARM code" and in fact there is no such thing as "Apple Silicon" code that is written by a developer (and again calling Apple Silicon ARM is not accurate). The basic difference between the current code and coding for ASi is mostly a recompile with a few cases of different security and/or addressing of some hardware.

I would classify any compiled binary that runs on ARM64 as "ARM Code". It is composed of ARM64 instructions. ARM64 assembly language is also "ARM Code" in my opinion because it only targets ARM processors.

Perhaps you meant application code in other languages for which there is an ARM compiler?

I would also argue that there is a concept of Apple Silicon Code if it uses Apple APIs that a specific to Apple Silicon. Apple is pretty good about providing device-independent APIs so this probably won't happen often, but I could conceive of apps that need to replace Intel-specific instructions (e.g. SSE & AVX) with the ARM equivalent, or some low-level instructions for a specific Apple SoC feature. It's an edge case for sure, and I would expect and hope that most well-written apps can just rebuild for ASi without any other changes.
[automerge]1600318091[/automerge]

 

[Yebubbleman]

As a video editor, this is where things get murky.

When most people (and probably you) talk about 'performance' you're probably talking about Geekbench scores (or something similar). But.... Geekbench scores are heavily influenced by tricks like hyper threading and turbo boosting.
Well, when you have any kind of long heavy duty process that could really benefit from that 'speed' ... guess what - the turbo boost benefit that shows up in Geekbench disappears because things like turbo boost are only meant to run for short periods of time (ESPECIALLY on the typically 'thin' apple products), and on processes that only require a core or two. But no processor can stably deliver long term performance at anywhere near peak turbo boost speeds.

So, all this bravado talk about how well Apple silicon will perform? That to me comes with that huge question mark: Apple silicon may beat Intel in a 2 minute benchmark, but how will it compare when I'm rendering a 1.5 hour video and the heat the pressure on the processors drop them back to their 'stable' performance levels?

To me, that is where the real proof in the pudding lies, and certainly nothing I've seen has given any sense of what the real answer will be.

So don't show me another stupid Geekbench score. Show me what sustained - long term - performance looks like when comparing the two processors. Until I see that, all you're talk about Apple silicon trouncing Intel chips just feels like vacuous slight of hand.

Not sure how long you've been using Macs or how familiar you are/were with the transition TO Intel processors that occurred between 2005 and 2006, but Apple is, so far, following the exact same playbook. The only really notable exception here is that, rather than producing a native version of the OS that was current at the time of the transition announcement for the new architecture, they're saying that it will start with the next release (i.e. Big Sur). That and they've announced that the first Mac will be announced so soon after the announcement (which can only signify that they're far along on this process). But otherwise, this is the same playbook from 2005-06.

Given that, it's important to note a few things:

- Apple will not EVER announce a new Mac that is slower than the Mac it replaces. It's simply not going to happen.

- Apple has SEVERAL skeptics (including you) who are unsure that this transition will actually result in FASTER performance, let alone "better products"; they are aware that they need to show proof or people will (a) not buy these new Macs and (b) cling to the Intel Macs that are out there which will screw them royally. This was true of the PowerPC to Intel transition and it is no less true of this Intel to Apple Silicon transition

- Apple only makes these kind of transitions when they hit a wall with the processor architecture they are currently on; with PowerPC, they couldn't get anything past a G4 on a notebook and they couldn't make a G5 that could go up to 3GHz and/or be even remotely power efficient and IBM (who made the PowerPC G5 series of processors) wasn't doing anything to help Apple on either front. With Intel right now, there has only been slight advancements since Skylake (6th Gen Intel CPUs that were on 2016 MacBook Pros and MacBooks, and 2015 Retina 5K 27" iMacs). The best we got was Intel increasing the core count with Coffee Lake (8th Gen). But past that, Intel has struggled to advance anything past how things were four years ago. When you talk about performance, as a video editor, it ought to worry you that Apple hasn't been able to offer anything drastically more powerful than the machine it was selling four years ago. Back in 2005, that would've been unheard of.

- If you look at the iPhone and iPad, there have been monumental leaps across certain A-series SoC generations. A8 to A9 was a monumental leap. A9 to A10 Fusion wasn't quite as substantial, but the introduction of asymmetric processing (high-efficiency cores + high performance cores) is huge. A10 Fusion to A11 Bionic was massive. And it's honestly been going from there. This is a processor family that has seriously steady growth in performance.


As for your comment that I'm only focusing on benchmarks and not real-world performance, I sort of take offense to that. (Especially since Apple has demonstrated native performance of Apple Silicon and shown that you can video edit on even a 2018 iPad Pro SoC [albeit with one extra GPU core enabled] based Mac mini and get far better performance than the current Intel one.) You don't get a 100% accurate view of how well something will perform based on Geekbench or XBench results. This is true, and I'll never argue against that. However, that's not the point of benchmarks. The point of benchmarks is to provide context and comparison between two different things to see just how far apart they are from each other. Is it 100% accurate representation? Again, no. Nor am I saying it is. But it ought to be SOME indication and what it's telling us is that the performance of Apple Silicon is going to be insane.

If a 2018 iPad Pro is holding its own being benchmarked against any Mac that isn't an i9 15" or 16" MacBook Pro, an 8-core or 10-core 27" iMac, an iMac Pro, or 2019 Mac Pro, then it's safe to say that, the native performance of an Apple Silicon SoC (which will undoubtedly be faster than said 2018 iPad Pro's SoC) will be enough to put all of the lower-end Intel Macs to shame right out of the gate. Furthermore, the 27" iMac getting an Intel-based update after the Apple Silicon announcement with the 21.5" iMac NOT getting ANY update is not unintentional. Clearly, Apple needs more time to get the performance down for the higher-end models. All that to say that, for you, a video editor, you're not going to see high-end Apple Silicon Macs this year for the same reason you are apprehensive about Apple Silicon to begin with; performance is not there yet. But you're blind if you don't see that they're going to get there before their two-year timeframe is up.

As for Turbo-Boost, the processor should be able to sustain the load by turning off the other cores. I'm not saying it's optimal. But most video apps are multi-core aware, meaning that Turbo-Boost ought to not be a concern. But, given that one needs to adhere to many other Apple technologies in porting to Apple Silicon anyway (i.e. shifting to Metal, adopting Grand Central Dispatch), this ought to not be a concern, unless you have an Apple dev that doesn't know how to develop for Apple. But, if you have someone developing for Apple platforms and is adhering to best practices when creating an Apple Silicon native macOS binary, then the results in real-world performance should be markedly better than what you get on Intel Macs.

I appreciate the skepticism (and certainly if it were my bread and butter, I'd be as well). But, honestly, it's gonna be okay and there's plenty of real world evidence to support that. Again, we went through this in 2005-06 and, if anything, they're doing a much smoother job of it this time around.

 

[Nicole1980]

Not sure how long you've been using Macs or how familiar you are/were with the transition TO Intel processors that occurred between 2005 and 2006, but Apple is, so far, following the exact same playbook. The only really notable exception here is that, rather than producing a native version of the OS that was current at the time of the transition announcement for the new architecture, they're saying that it will start with the next release (i.e. Big Sur). That and they've announced that the first Mac will be announced so soon after the announcement (which can only signify that they're far along on this process). But otherwise, this is the same playbook from 2005-06.

Given that, it's important to note a few things:

- Apple will not EVER announce a new Mac that is slower than the Mac it replaces. It's simply not going to happen.

- Apple has SEVERAL skeptics (including you) who are unsure that this transition will actually result in FASTER performance, let alone "better products"; they are aware that they need to show proof or people will (a) not buy these new Macs and (b) cling to the Intel Macs that are out there which will screw them royally. This was true of the PowerPC to Intel transition and it is no less true of this Intel to Apple Silicon transition

- Apple only makes these kind of transitions when they hit a wall with the processor architecture they are currently on; with PowerPC, they couldn't get anything past a G4 on a notebook and they couldn't make a G5 that could go up to 3GHz and/or be even remotely power efficient and IBM (who made the PowerPC G5 series of processors) wasn't doing anything to help Apple on either front. With Intel right now, there has only been slight advancements since Skylake (6th Gen Intel CPUs that were on 2016 MacBook Pros and MacBooks, and 2015 Retina 5K 27" iMacs). The best we got was Intel increasing the core count with Coffee Lake (8th Gen). But past that, Intel has struggled to advance anything past how things were four years ago. When you talk about performance, as a video editor, it ought to worry you that Apple hasn't been able to offer anything drastically more powerful than the machine it was selling four years ago. Back in 2005, that would've been unheard of.

- If you look at the iPhone and iPad, there have been monumental leaps across certain A-series SoC generations. A8 to A9 was a monumental leap. A9 to A10 Fusion wasn't quite as substantial, but the introduction of asymmetric processing (high-efficiency cores + high performance cores) is huge. A10 Fusion to A11 Bionic was massive. And it's honestly been going from there. This is a processor family that has seriously steady growth in performance.


As for your comment that I'm only focusing on benchmarks and not real-world performance, I sort of take offense to that. (Especially since Apple has demonstrated native performance of Apple Silicon and shown that you can video edit on even a 2018 iPad Pro SoC [albeit with one extra GPU core enabled] based Mac mini and get far better performance than the current Intel one.) You don't get a 100% accurate view of how well something will perform based on Geekbench or XBench results. This is true, and I'll never argue against that. However, that's not the point of benchmarks. The point of benchmarks is to provide context and comparison between two different things to see just how far apart they are from each other. Is it 100% accurate representation? Again, no. Nor am I saying it is. But it ought to be SOME indication and what it's telling us is that the performance of Apple Silicon is going to be insane.

If a 2018 iPad Pro is holding its own being benchmarked against any Mac that isn't an i9 15" or 16" MacBook Pro, an 8-core or 10-core 27" iMac, an iMac Pro, or 2019 Mac Pro, then it's safe to say that, the native performance of an Apple Silicon SoC (which will undoubtedly be faster than said 2018 iPad Pro's SoC) will be enough to put all of the lower-end Intel Macs to shame right out of the gate. Furthermore, the 27" iMac getting an Intel-based update after the Apple Silicon announcement with the 21.5" iMac NOT getting ANY update is not unintentional. Clearly, Apple needs more time to get the performance down for the higher-end models. All that to say that, for you, a video editor, you're not going to see high-end Apple Silicon Macs this year for the same reason you are apprehensive about Apple Silicon to begin with; performance is not there yet. But you're blind if you don't see that they're going to get there before their two-year timeframe is up.

As for Turbo-Boost, the processor should be able to sustain the load by turning off the other cores. I'm not saying it's optimal. But most video apps are multi-core aware, meaning that Turbo-Boost ought to not be a concern. But, given that one needs to adhere to many other Apple technologies in porting to Apple Silicon anyway (i.e. shifting to Metal, adopting Grand Central Dispatch), this ought to not be a concern, unless you have an Apple dev that doesn't know how to develop for Apple. But, if you have someone developing for Apple platforms and is adhering to best practices when creating an Apple Silicon native macOS binary, then the results in real-world performance should be markedly better than what you get on Intel Macs.

I appreciate the skepticism (and certainly if it were my bread and butter, I'd be as well). But, honestly, it's gonna be okay and there's plenty of real world evidence to support that. Again, we went through this in 2005-06 and, if anything, they're doing a much smoother job of it this time around.

Again, you say Apple chips will be 'faster', based on bench and Geekbench. You also add that the performance gain will be 'INSANE". Based on what metric do you get INSANE?

So I take issue with Geekbench and xBench but you reference them again, and what 'INSANE' leap are you looking at? From what I can tell, some Apple chips may be a tad faster in single core, and then somewhat slower in multi core (using those benchmarks I take issue with anyways).

The move to 3nm process will help apple chips a little, but it won't be 'INSANE'. And then I ask you, how long can TMSC keep shrinking something that's already down to 3nm? I guess we'll find out.

And no, Turbo Boost does not sustain when all cores are being pushed for more than a few minutes, and it certainly won't sustain in a thin apple product that's already challenged to dissipate heat. At least, that's my opinion.

Like I've said, I'll be real interested to see what performance ends up being when all the cores are pushed pretty hard ... given what I do, that's the only situation where all that extra supposed 'power' will matter. If it can't deliver that, then all your 2-minute benchmarking tools are completely useless to me.

 

[Joe Dohn]

So, all this bravado talk about how well Apple silicon will perform? That to me comes with that huge question mark: Apple silicon may beat Intel in a 2 minute benchmark, but how will it compare when I'm rendering a 1.5 hour video and the heat the pressure on the processors drop them back to their 'stable' performance levels?

Oh,video rendering will probably be amazing, considering that Apple has specialized chips for video rendering even on their phones (why else do you think the iPhone can get that 4k, 60 FPS recording reliably)?

The problem is that video rendering is only a small fraction of the whole market. Unless performance with Apple Silicon is very extraordinary (and not "just" 20% to 30% faster), I don't see how the average user will switch from x86 to Apple Silicon. And I really don't think it is, or else Apple would be gloating over it from the start.

 

[leman]

Like I've said, I'll be real interested to see what performance ends up being when all the cores are pushed pretty hard ... given what I do, that's the only situation where all that extra supposed 'power' will matter. If it can't deliver that, then all your 2-minute benchmarking tools are completely useless to me.

I think the point that you might be missing is the power operating range. A single core in a current Intel CPU can draw somewhere between 3 watts and 25-30 watts in typical operation. It's because Intel's clock frequency range is insanely wide: from 1.3 ghz (low power state) to up to 5 ghz or higher.

Apple CPUs frequency range — and the peak clocks — are much lower in comparison, and their power curve is very aggressive. The max turbo boost for an A13 core is just 2.7ghz, with CPU power draw around 5 watts. But it draws only half that power when you run it at 2.5ghz for example.

What this means in practice is that you can get a large improvement in power efficiency of Apple CPUs by slowing them down only marginally. With Intel, you need to apply a much more substantial speed reduction to maintain a target power figure. To put it into perspective, 8 Apple A13 cores running at 2.7ghz would meet the TDP of 40 watts. Looking at my 16" MBP, 8 Intel cores running at 3.2ghz draw close to 60 watts. Given the fact that an A13@2.7 ghz is roughly equivalent in performance to an Intel core at 4.5ghz+, Apple has an undeniable advantage.

In fact, looking at the available data I have little doubt that Apple will dramatically outperform intel in sustained operation. Given that their cores do much more work per clock, this is exactly their forte. The big unknown is the peak performance — will Apple manage to break the current single-threaded plateau that neither Intel nor AMD have been able to cross?

 

[fokmik]

Ok, this is what the upcoming arm based Macbook Pro 14" must top ( and not the current one, the current 13" Mbp is an easy task for apple, even the dev kit mac mini is already on top of it) :

 

[Nicole1980]

Ok, this is what the upcoming arm based Macbook Pro 14" must top ( and not the current one, the current 13" Mbp is an easy task for apple, even the dev kit mac mini is already on top of it) :

This is what would potentially crack me up - if after the switch to apple silicon, intel chips end up leapfroging ahead of apple chips and all the fanboys will be looking around confused saying 'what the heck just happened'?!?

 

[Jorbanead]

In the short term, I think I am talking about what a Mac can and can't do using x86 versus what its limits are with Apple Silicon. The potential growth of either system I couldn't say, but with Intel/AMD x86 the potential for massive improvement isn't there, and probably is for an ARM chip. But right now software is more advanced and functional for the x86. And I don't know what limits an ARM has or doesn't have in comparison or how long until the software catches up. Thats not just a hardware situation its a who is developing what and how hard is it to write ARM code to do what x86 code can already do.

I'm not a software dev so someone else may correct me, but from my understanding if your software is written in swift UI or you're using Xcode, the transition is just a simple recompile and that's it. Even older code like photoshop (which is notorious for having ancient code) only took adobe a few days to recompile (From Craig in an interview).

 

[raknor]

This is what would potentially crack me up - if after the switch to apple silicon, intel chips end up leapfroging ahead of apple chips and all the fanboys will be looking around confused saying 'what the heck just happened'?!?

https://www.anandtech.com/show/16084/intel-tiger-lake-review-deep-dive-core-11th-gen/8

The i7-1185G7 is at the heels of the desktop i9-10900K, trailing only by a few percentage points.

Against the x86 competition, Tiger Lake leaves AMD’s Zen2-based Renoir in the dust when it comes to single-threaded performance. Comparing it against Apple’s A13, things aren’t looking so rosy as the Intel CPU barely outmatches it even though it uses several times more power, which doesn’t bode well for Intel once Apple releases its “Apple Silicon” Macbooks.

Even against Arm’s Cortex-A77 things aren’t looking rosy, as the x86 crowd just all that much ahead considering the Arm design only uses 2W.

Doesn’t look like that is happening this generation. Even ARM cores in the graviton are doing quite well with lower powers. The next gen ARM reference cores have a nice performance uplift. Things are looking good for ARM to make headway into PCs and even the data centers in the coming years outside of Apple’s ecosystem. The Intel fanboys might be the ones left wondering....

 

[Krevnik]

I guess Microsoft is insane then. The Mac BU is a separate entity from the Windows Office variant and do not share code at the level you think they do.

They’ve since re-merged the codebases according to Erik Schwiebert‘s thread:

https://twitter.com/i/web/status/954037656677072896

Erik also gave a talk on this that gets into more details on it:

 

[leman]

Regarding Tiger Lake. Anandtech review is out, with the following results:

- CPU-wise, Tiger Lake at 28 watts (with the CPU actually running at 50 Watts most of the time) is barely faster than an iPhone A13 at 5 watts

- GPU-wise, it's a bit more difficult to estimate. Intel's improvements in GPU performance is absolutely remarkable, the team did a very good job (them hiring former AMD's and Apple's graphics hardware guru Raja Koduri has apparently paid off). As a ballpark estimate, I'd say that in the 28W TDP, Tiger Lake 96EU Xe-LP is around 50-60% faster than the A12Z. Apple GPU still seems to have about 2x advantage in performance-per-watt.

 

[EdT]

I never owned a PowerPC Mac, which was a ARM processor. I didn’t because I used a desktop and not a laptop and battery life and heat were completely unimportant to me. And today for my Desktop I still don’t care if it uses more watts of power and generates more heat. If I have a cheap laptop then I might exchange computational power for battery life, but not if I owned a MacBook Pro.

I get that Apple is having problems with Intel getting new processors out in a timely fashion, and that x86 chips are one big reason that battery life is stuck at around 6-10 hours for anyone’s brand of medium or better performance laptop. But I haven’t heard WHY or WHAT has changed in ARM processors that makes them attractive now to (computational) power users. That’s more or less the original point of the person that started this post: Many people don’t want a computer that is just a big iPad. My iPad 2019 Air doesn’t run Lightroom or Photoshop as quickly as my current desktop, which is a 5 year old iMac. I don’t really care what the processor guts are I care what the program outcome is.

 

[jdb8167]

I never owned a PowerPC Mac, which was a ARM processor. I didn’t because I used a desktop and not a laptop and battery life and heat were completely unimportant to me. And today for my Desktop I still don’t care if it uses more watts of power and generates more heat. If I have a cheap laptop then I might exchange computational power for battery life, but not if I owned a MacBook Pro.

I get that Apple is having problems with Intel getting new processors out in a timely fashion, and that x86 chips are one big reason that battery life is stuck at around 6-10 hours for anyone’s brand of medium or better performance laptop. But I haven’t heard WHY or WHAT has changed in ARM processors that makes them attractive now to (computational) power users. That’s more or less the original point of the person that started this post: Many people don’t want a computer that is just a big iPad. My iPad 2019 Air doesn’t run Lightroom or Photoshop as quickly as my current desktop, which is a 5 year old iMac. I don’t really care what the processor guts are I care what the program outcome is.

PowerPC was not ARM but it’s own RISC architecture from IBM.

You might not care about power and heat if you are only using desktops but if you use a notebook you definitely care because of battery life and performance.

Your 2019 iPad Air runs a CPU from 2018—the A12. It’s a decently capable CPU but it will be leapfrogged by whatever Apple is calling the first Apple Silicon CPU. The A12 has 2 high performance cores/4 efficiency cores for the CPU and only 4 GPU cores. It was the low end from Apple in 2018. The A12Z in the 2020 iPad Pro is 4 performance/4 efficiency CPU and 8 GPU cores. This is a CPU also designed in 2018. So even in the same generation, the A12Z is almost twice as fast as the A12.

The new A14 in the iPad Air is using a CPU feature size that is the best in the world. There is no debate from anyone knowledgeable about whether TSMC’s 5 nm is the best that can be done currently. Your programs are going to run several times faster on a Mac with Apple Silicon than what you see on your old iPad Air.

 

[Jorbanead]

But I haven’t heard WHY or WHAT has changed in ARM processors that makes them attractive now to (computational) power users.

Many reasons to be excited:

1. Apple Silicon already beats out intel in single-core performance, which has gone stagnant for intel
2. Better performance overall (Using a smaller manufacturing process) with 5nm and soon to be 3nm in a few years. Intel is still stuck on 10nm (7nm TMSC Equivalent). Apple can be on the leading edge of the new node process using their own chips.
3. Tighter software/hardware integration. If you use Final Cut Pro or Logic Pro, AS Macs are going to be a game-changer because: Hardware encoders, decoders, and co-processors to handle audio and video encoding and DSP capabilities, and access to machine learning. Apple can offer features in these applications that utilize special hardware on their silicon that could surpass anything on an intel chip.
4. bigLITTLE allows power users to offload background tasks and rendering to the efficient cores, while you utilize the performance cores for your active application

 

[ericwn]

This is what would potentially crack me up - if after the switch to apple silicon, intel chips end up leapfroging ahead of apple chips and all the fanboys will be looking around confused saying 'what the heck just happened'?!?

In return let’s have the haters apologize if Apple silicone is indeed a vast improvement over the current intel stuff.

 

[dinobear]

Ok, this is what the upcoming arm based Macbook Pro 14" must top ( and not the current one, the current 13" Mbp is an easy task for apple, even the dev kit mac mini is already on top of it) :

This is why Apple won't upgrade thier macbooks to intels new chips. They want to advertise the new ARM macbook to be X-times faster than the "previous model". The previous model being an 8th gen intel 😂

 

[Joe Dohn]

This is what would potentially crack me up - if after the switch to apple silicon, intel chips end up leapfroging ahead of apple chips and all the fanboys will be looking around confused saying 'what the heck just happened'?!?

This is a very real possibility, especially considering Apple doesn't refresh its line as often.