Leaving Apple over ARM Based Macs discussion [merged]

[Breaking Good]

If you're referring to Microsoft and Adobe not wanting to port to ARM, and you know they won't...

I don't understand what you are saying here. Windows 10 has been running on Qualcomm Snapdragon chips since the Snapdragon 835. They integrate in with Qualcomm's modems. The latest is the Snapdragon 8cx and is supposed to allow 5G connectivity. It is part of the Windows Always Connected program.

Unfortunately, they haven't gotten much traction. I personally blame computer reviewers because they want to compare them to Intel Core i5 processors even though the vast majority of people using Windows PC's don't need the processing power of a Core i5 processor.

Hopefully, Apple moving Mac's to ARM will give Windows on Snapdragon (WoS) a boost.

 

[SocialKonstruct]

I don't understand what you are saying here. Windows 10 has been running on Qualcomm Snapdragon chips since the Snapdragon 835. They integrate in with Qualcomm's modems. The latest is the Snapdragon 8cx and is supposed to allow 5G connectivity. It is part of the Windows Always Connected program.

Unfortunately, they haven't gotten much traction. I personally blame computer reviewers because they want to compare them to Intel Core i5 processors even though the vast majority of people using Windows PC's don't need the processing power of a Core i5 processor.

Hopefully, Apple moving Mac's to ARM will give Windows on Snapdragon (WoS) a boost.

Unfortunately installing Steam on the Microsoft Surface X was a dismal experience. The X went back to where it failed to mark the spot.
[automerge]1591840447[/automerge]

The day they don't allow third party software (ala iOS) is the day I start looking to leave.

I have a full blown computer for a reason -- and having my experience "curated" (neutered) by the manufacturer is of no interest to me.
[automerge]1591831616[/automerge]


I think the skepticism here is because Apple is horrible at software basically anytime in recent memory.

The hardware I have faith in -- but the more "iOS like" the lockdown and software restrictions become on the macOS side, the less desirable the platform becomes (at least for me).

"Lockdown" is tolerated on an iPhone...less enjoyable on the iPad and, to me, totally unacceptable, on a full blown main computing device.

And that worries me a lot about Apple going that direction. If I am forced to use the App Store to install stuff on a pro machine then I am definitely keep my Macs for legacy devices only and switching to Surface Book full time.

 

[Darth Tulhu]

The day they don't allow third party software (ala iOS) is the day I start looking to leave.

I have a full blown computer for a reason -- and having my experience "curated" (neutered) by the manufacturer is of no interest to me.
[automerge]1591831616[/automerge]


I think the skepticism here is because Apple is horrible at software basically anytime in recent memory.

The hardware I have faith in -- but the more "iOS like" the lockdown and software restrictions become on the macOS side, the less desirable the platform becomes (at least for me).

"Lockdown" is tolerated on an iPhone...less enjoyable on the iPad and, to me, totally unacceptable, on a full blown main computing device.

Understandable sentiments, even though to me they seem premature.

Wouldn't you want to wait and see if the App Store meets your pro needs before you jump ship?

Personally, I'd do ANYTHING to not have to use Windows as my main device. But I understand business is business, so hopefully Apple won't drop the ball here, and so far they seem to be paying attention to the Pros recently.

 

[sixtydashone]

They won't switch completely any time soon, but when they do, as long as the ARM machines are stable, acceptably powerful on a true desktop OS, and compatible with all the software I need them to be compatible with, then I don't see any reason to move away from them.

 

[lJoSquaredl]

I use FCPX so I'd just upgrade and enjoy the quiet cool running computer I use mostly native MacOS apps anyways, and I'm wired in pretty tight with iCloud stuff, I'd maybe have to something Windows based for gaming but eh...been missing my gaming PC setup anyways. This would just be a good excuse to have the MBP/Gaming PC combo again.

 

[UBS28]

This thread highlights the two main scenarios regarding the Mac switching to ARM.

I can't wait to come back to it once we ACTUALLY know what Apple is going to do.

I think Apple is being underestimated though, even if Steve is no longer around to drive the transition as smoothly as before.

I would love to see Apple pull out a SERIOUS A-chip that blows x86 out of the water, and makes macOS great again.

The macOS is just too good for it to just go away, so I think those worried it'll be replaced with iPadOS are overreacting.

This WWDC will be very, very interesting.

ARM does not blow away X86 out of the water. It is simply a reduced instruction set which allows it to have way less transistors than X86. That is how it is able to achieve such a low power for mobile devices. X86 has complex instructions which are not needed for simple tasks that mobile devices do, so that is why ARM is more suitable in an iPhone or iPad (as the most complex X86 instruction set will not be used and cause extra battery drain).

For devices were complex tasks are done, that is were ARM will start to struggle as it requires much more operations than a X86 processor. X86 also has much more compatibility and is able to optimize incoming instructions.

What Apple should do is switch to AMD instead of ARM for their pro devices and do ARM for devices that do simple tasks. An AMD EPYC has no problem keeping up with a 80-core ARM server cpu, so there is no need to go to ARM for a MAC pro for example right now (maybe in a few years when all software has been ported to ARM)

 

[bill-p]

The main point of Catalina was to drop support for 32-bit apps. I’m sure Apple’s plan is to make it relatively easy to port 64-bit apps from Intel to ARM.

The problem is not just about 32-bit or 64-bit. I think I mentioned this in another thread but there are a lot of legacy Java, Electron (heck, Electron is NOT legacy), and etc... other apps that will need to be completely reworked for this new ARM Mac. Some of these are just a matter of developers going back and dusting off their old code base, but a lot of these come down to the fact that code is dead or no longer maintained and there is no easy way to bring them to ARM.

So then a lot of apps will need to be rewritten (some from scratch) if x86 support is abruptly cut. That's detrimental to consumers.

There's a reason the Surface Pro X failed. It failed even with x86 emulation in tow. I'm watching intently to see how Apple will solve the problem but... honestly, it's not going to be an overnight thing. It's not even going to be something that can be achieved within a year.

Let's just say... I understand the excitement, but you guys are way overestimating what is possible with ARM processors.

... so far they seem to be paying attention to the Pros recently.

I'm so very sorry, but... HAHAHAHAHAHAHAHAHAHA!

HAH!

HAH!

If Apple was truly listening to Pros, they would:

1. Ditch Touch Bar (or make it optional)
2. Give us more than USB-C style ports
3. Give us nVidia instead of AMD
4. Not force Metal down our throats, give us Vulkan.
5. Release a MacOS version that's at least functional without major breaking bugs.

And I'm sure everyone in this thread knows that none of those things will happen. Apple has slipped a lot recently. Their hardware is the only thing that's still solid, but that's debatable even.

 

[KPOM]

The problem is not just about 32-bit or 64-bit. I think I mentioned this in another thread but there are a lot of legacy Java, Electron (heck, Electron is NOT legacy), and etc... other apps that will need to be completely reworked for this new ARM Mac. Some of these are just a matter of developers going back and dusting off their old code base, but a lot of these come down to the fact that code is dead or no longer maintained and there is no easy way to bring them to ARM.

So then a lot of apps will need to be rewritten (some from scratch) if x86 support is abruptly cut. That's detrimental to consumers.

There's a reason the Surface Pro X failed. It failed even with x86 emulation in tow. I'm watching intently to see how Apple will solve the problem but... honestly, it's not going to be an overnight thing. It's not even going to be something that can be achieved within a year.

Let's just say... I understand the excitement, but you guys are way overestimating what is possible with ARM processors.



I'm so very sorry, but... HAHAHAHAHAHAHAHAHAHA!

HAH!

HAH!

If Apple was truly listening to Pros, they would:

1. Ditch Touch Bar (or make it optional)
2. Give us more than USB-C style ports
3. Give us nVidia instead of AMD
4. Not force Metal down our throats, give us Vulkan.
5. Release a MacOS version that's at least functional without major breaking bugs.

And I'm sure everyone in this thread knows that none of those things will happen. Apple has slipped a lot recently. Their hardware is the only thing that's still solid, but that's debatable even.

Java and Electron apps are dinosaurs. Plain and simple. They are garbage.

As for USB-C, it is the modern do-all port. There is nothing “pro” about legacy ports. TB3 can connect to every legacy device with the right adapter. You can’t say that about any other port. As for Metal, use it. It’s clear that Apple is looking ahead to cross platform functionality between iOS and macOS.

 

[bill-p]

Java and Electron apps are dinosaurs. Plain and simple. They are garbage.

As for USB-C, it is the modern do-all port. There is nothing “pro” about legacy ports. TB3 can connect to every legacy device with the right adapter. You can’t say that about any other port. As for Metal, use it. It’s clear that Apple is looking ahead to cross platform functionality between iOS and macOS.

There are a lot of Java and Electron apps still operational in the "pro" world, whether you like it or not.

Prime example? Visual Studio Code. Good luck trying to convince us developers that VS Code is "garbage".

And the problem with USB-C is not its utility. It's the fact that it's not the only thing in existence. There are still USB-A and SD cards and Ethernet cables that do not plug directly into USB-C. I've lived with dongles for 5 years now, hoping the world will move on the next year, but the fact is: it won't. Apple is only limiting its users, and it's clear Apple isn't influential enough with the push for USB-C.

Hell, count the number of Thunderbolt 3 displays on the market. I'm sure you'll be surprised.

 

[KPOM]

There are a lot of Java and Electron apps still operational in the "pro" world, whether you like it or not.

Prime example? Visual Studio Code. Good luck trying to convince us developers that VS Code is "garbage".

And the problem with USB-C is not its utility. It's the fact that it's not the only thing in existence. There are still USB-A and SD cards and Ethernet cables that do not plug directly into USB-C. I've lived with dongles for 5 years now, hoping the world will move on the next year, but the fact is: it won't. Apple is only limiting its users, and it's clear Apple isn't influential enough with the push for USB-C.

Hell, count the number of Thunderbolt 3 displays on the market. I'm sure you'll be surprised.

Java and electron are bad. They are bloated and inefficient.

I’ve been exclusively USB-C for years now. I use a Thunderbolt dock for my work PC that handles any and all legacy connections the rare occasions I need them. I don’t miss any legacy port. My main external display is the LG Ultrafine 5K. Its gorgeous screen works with both my Windows PC and Mac through the Thunderbolt ports in both.

 

[AndyMacAndMic]

Java and electron are bad. They are bloated and inefficient.

I’ve been exclusively USB-C for years now. I use a Thunderbolt dock for my work PC that handles any and all legacy connections the rare occasions I need them. I don’t miss any legacy port. My main external display is the LG Ultrafine 5K. Its gorgeous screen works with both my Windows PC and Mac through the Thunderbolt ports in both.

'Bad' is a useless qualification. As long as Java etc. are extensively used in the professional world they are needed by professionals, 'bad' or not.

Your anecdotal example about USB-C does not speak for millions of other professionals in the world who don't share your experience.

Things are not as black and white as your oversimplification makes them appear.

 

[bluecoast]

Yes, if I was using an key application for work that was Windows Intel (64 bit Win32) only and I wanted to get an ARM Mac. And maybe if I wanted a gaming rig.

Otherwise, no, because Windows is a mess.

However in 2021, if Windows X launches with a consistent UI and the old legacy stuff ripped out, it could be a great draw for consumers.

Ditto for work, if the virtualisation & fast OS updates etc plus a WinUI makeover for legacy Win32 style apps can be applied to Windows 10, then Windows could be looking pretty good in 2021.

Trouble is, that MS have a track record of making noises about cleaning up Windows which they never quite do.

I think that Windows XP was the closest that they ever got to having a consistent UI experience in Windows this century (although some obscure utilities were still using Windows 3.11 style open/save dialogs).

 

[KPOM]

'Bad' is a useless qualification. As long as Java etc. are extensively used in the professional world they are needed by professionals, 'bad' or not.

Your anecdotal example about USB-C does not speak for millions of other professionals in the world who don't share your experience.

Things are not as black and white as your oversimplification makes them appear.

Rene Ritchie has written extensively about how bloated and inefficient Electron apps are. A big part of Microsoft’s problems stem from its fanatical devotion to backward compatibility.

As for USB-C, it’s 2020. We should be almost exclusively USB-C by now. In 2003 (5 years after Apple released the iMac with no legacy ports) no one was complaining that “pros” needed SCSI or ADB ports.
[automerge]1591861539[/automerge]

Yes, if I was using an key application for work that was Windows Intel (64 bit Win32) only and I wanted to get an ARM Mac. And maybe if I wanted a gaming rig.

Otherwise, no, because Windows is a mess.

However in 2021, if Windows X launches with a consistent UI and the old legacy stuff ripped out, it could be a great draw for consumers.

Ditto for work, if the virtualisation & fast OS updates etc plus a WinUI makeover for legacy Win32 style apps can be applied to Windows 10, then Windows could be looking pretty good in 2021.

Trouble is, that MS have a track record of making noises about cleaning up Windows which they never quite do.

I think that Windows XP was the closest that they ever got to having a consistent UI experience in Windows this century (although some obscure utilities were still using Windows 3.11 style open/save dialogs).

Windows 10 isn’t a bad OS overall. However, MS has not been successful at getting developers to migrate to x64. Heck, Office still has a “32-bit vs 64-bit” flowchart because of legacy add-ins.” I’m sure Microsoft is rooting for Apple’s success with ARM since they would like to be free of legacy code, too. If Apple succeeds it gives Microsoft some leverage to tell their OEM customers to get with the program.

 

[AndyMacAndMic]

Rene Ritchie has written extensively about how bloated and inefficient Electron apps are.

Because Rene Ritchie writes about bloated Electron apps it does not take away from the fact that lots of professionals are using Java and Electron right now. What I said: It doesn't matter if something is 'bloated' or 'Bad' per your qualification. If it is used by millions of professionals it is still very relevant for professional use.

As for USB-C, it’s 2020. We should be almost exclusively USB-C by now.

'Should' is the keyword here, but in practice in 2020 we aren't, are we?

 

[Darth Tulhu]

I'm so very sorry, but... HAHAHAHAHAHAHAHAHAHA!

HAH!

HAH!

If Apple was truly listening to Pros, they would:

1. Ditch Touch Bar (or make it optional)
2. Give us more than USB-C style ports
3. Give us nVidia instead of AMD
4. Not force Metal down our throats, give us Vulkan.
5. Release a MacOS version that's at least functional without major breaking bugs.

And I'm sure everyone in this thread knows that none of those things will happen. Apple has slipped a lot recently. Their hardware is the only thing that's still solid, but that's debatable even.

That's fair, although I think there is a shift going on. I can see a difference in direction from 2016, but I think it'll take time. It looks like they're trying.

I think the new Mac Pro (even if it's still a bit out of touch), the 16" MBP, giving the iPad mouse support, and finally the killing off of the BS butterfly keyboard are evidence of this.

Now, from your list only 1 and 5 are possible. 2-4 ain't happening, so we'll all have to come to peace with this.
[automerge]1591876905[/automerge]

ARM does not blow away X86 out of the water. It is simply a reduced instruction set which allows it to have way less transistors than X86. That is how it is able to achieve such a low power for mobile devices. X86 has complex instructions which are not needed for simple tasks that mobile devices do, so that is why ARM is more suitable in an iPhone or iPad (as the most complex X86 instruction set will not be used and cause extra battery drain).

For devices were complex tasks are done, that is were ARM will start to struggle as it requires much more operations than a X86 processor. X86 also has much more compatibility and is able to optimize incoming instructions.

What Apple should do is switch to AMD instead of ARM for their pro devices and do ARM for devices that do simple tasks. An AMD EPYC has no problem keeping up with a 80-core ARM server cpu, so there is no need to go to ARM for a MAC pro for example right now (maybe in a few years when all software has been ported to ARM)

Fair enough, but from what I understand the A-Series chips are custom, so maybe Apple has something up their sleeve?

I honestly don't know and I'm nowhere near a hardware engineer. I guess I'm just a little overexcited about the possibilities and can't wait to see if, IF Apple delivers something surprising and delightful, like the days of old.

 

[leman]

ARM does not blow away X86 out of the water. It is simply a reduced instruction set which allows it to have way less transistors than X86. That is how it is able to achieve such a low power for mobile devices. X86 has complex instructions which are not needed for simple tasks that mobile devices do, so that is why ARM is more suitable in an iPhone or iPad (as the most complex X86 instruction set will not be used and cause extra battery drain).

For devices were complex tasks are done, that is were ARM will start to struggle as it requires much more operations than a X86 processor. X86 also has much more compatibility and is able to optimize incoming instructions.

This is not entirely correct. Yes, ARM sometimes needs more instructions to encode the same operation than x86-64... at the same time ARM has shorter instructions (compared to "complex" x86 instructions) and more registers. Practical comparisons of binary code size shows no substantial difference between ARM and x86... in many cases ARM actually produces denser code (as an example, here is a trivial quick sort implementation for 64bit ARM vs 64bit Intel: https://godbolt.org/z/86QLQ4 Aarch64 binary is 128 bytes, x86-64 binary is 192 bytes).

The "reduced instruction set" of ARM primarily allows it to save transistors for the decode logic. The CPUs themselves can be made as simple or as complex as the chip designers want. Modern high-performance ARM chips work very similarly to the x86 chips — both break down the instructions to CPU-specific micro-instructions, both reorder the instructions and aggressively preload data to maximize performance. Neither instruction set is inherently faster than the other one, but as mentioned above, ARM is simpler to decode, which potentially allows the chip designers to use the transistor budget for something mr useful (e.g. cache or additional execution units).

What do you mean when you say that x86 "has more compatibility"? Compatibility to what?

I am also not sure what you mean with "complex" or "simple" tasks. For all intends and purposes mobile devises are general-purpose computers and have to run the same kind of general-purpose software as any other machine. For pro-level workloads, I'd say that ARM ISA is superior, since they offer future-proof scalable vector instructions and a rich set of operations. With x86, you need to know the CPU you are targeting so that you can generate the fastest possible vectorized code. With ARM's SVE, the code will automatically take advantage of whatever the hardware can do.

 

[jmoore5196]

I'm not sure I understand the question. What "Professional" work are you doing that requires a Microsoft device?

I like to think I've added 6 or 7 years to my life expectancy by never having owned a Microsoft product. I'm not sure I'd be willing to obliterate that kind of time benefit unless Adobe, et al, were unwilling to develop more specialized apps for ARM.

 

[jjjoseph]

If Apple goes ARM completely, they're going to have worked with Adobe and Microsoft and all of the other main players to make sure that software isn't running slower...

Adobe and Microsoft are not exactly BFF's with Apple. I honestly hope Apple does software again. Final Cut Pro, Apple Shake, Aperture, DVD Studio Pro.. They were all highly productive great apps that made the Apple walled garden great. If they go to ARM, my hope is for that kind of renaissance.

 

[petvas]

Adobe and Microsoft are not exactly BFF's with Apple. I honestly hope Apple does software again. Final Cut Pro, Apple Shake, Aperture, DVD Studio Pro.. They were all highly productive great apps that made the Apple walled garden great. If they go to ARM, my hope is for that kind of renaissance.

This will not be enough for many pros out there. Microsoft Office and the Adobe Creative Suite are two of the most used productivity tools and they must become available for ARM processors when Apple makes the switch.

 

[thejoshhoward]

Adobe and Microsoft are not exactly BFF's with Apple. I honestly hope Apple does software again. Final Cut Pro, Apple Shake, Aperture, DVD Studio Pro.. They were all highly productive great apps that made the Apple walled garden great. If they go to ARM, my hope is for that kind of renaissance.

I miss Aperture so much.

 

[Falhófnir]

The iPad Pro has 8 cores, so the 12 core ARM cpu should be fine in the MBA since it has better cooling than the iPad Pro.

I don't think a professional machine can be switched to ARM right now. But let's wait and see what happens.

This is exactly why I think Apple will push out powerful Arm Macs from the get go, this is not the idea of Arm Macs they want to get into people's heads if they are going to be switching the MacBook Pro, Mac Pro and iMac Pro eventually. They need big and impressive from the start to break down this idea people have of low power, low performance chips.

 

[Xenobius]

If Apple goes ARM it will leave Intel crap machines in the dust completely...

 

[bill-p]

Or they will push out a disappointing midrange ARM chip that can't even match a Core i5.

Then you guys will see that the reality is ARM chips are only good for low power, low performance tasks.

This is not just a theory. Cortex-A15 has already been tested to actually not be more power efficient than x86 given a sufficiently complex workload:

This is the data that shoots down the “ARM is intrinsically more power efficient than x86” argument. In mobile, where ARM is strongest, the Clover Trail Atom is more efficient than the Cortex-A8, but doesn’t quite match the Cortex-A9. Bobcat is actually less efficient than the Core i7, and the Cortex-A15 draws far more power than any previous ARM designs.

Source: https://www.extremetech.com/extreme...-or-mips-intrinsically-more-power-efficient/2

The move to ARM will only allow Apple more control over their hardware platforms, and also more profit margins. But in terms of actual performance, I'm not holding my breath until Apple can prove to me they can design an ARM chip that outperforms a Core i5 at the very least.

And I'm pretty sure someone in here will cite Geekbench, but Geekbench can hardly be qualified as a "complex" workload.

 

[oneMadRssn]

The move to ARM will only allow Apple more control over their hardware platforms, and also more profit margins.

I don't know if I buy this premise.

If Apple wanted more control and lower component costs, while maintaining maximum compatibility, they would switch to AMD Ryzen 4000 chips. I'm sure AMD would even make them custom skus to fit their needs, and they would be fabbed at Apple's fab partner TSMC using their 7nm process node (or upcoming 5nm). This would get them almost the same thing as switching to ARM (chips for iDevices and Macs rolling off the same assembly line at TSMC), but with a lot less headache for everyone else because it would maintain x64 compatibility.

In terms of control, sure Apple gets to dictate the exact layout of their A-series chips. But they're still beholden to TSMC's advances in process node technology. Today TSMC is leading but maybe tomorrow UMC or Samsung or GlobalFoundries will be leading, and Apple can easily switch from one to another - something they can't do with Intel. But they can do the same switching from fab to fab while using AMD.

In terms of cost, AMD costs much less than Intel. And I'm sure they would love to work with Apple and would create custom models or skus for them.

So if Apple is switching Macs to ARM, there has to be more to it than just hardware flexibility and profit margins.

 

[sub150]

Or they will push out a disappointing midrange ARM chip that can't even match a Core i5.

Then you guys will see that the reality is ARM chips are only good for low power, low performance tasks.

This is not just a theory. Cortex-A15 has already been tested to actually not be more power efficient than x86 given a sufficiently complex workload:



Source: https://www.extremetech.com/extreme...-or-mips-intrinsically-more-power-efficient/2

The move to ARM will only allow Apple more control over their hardware platforms, and also more profit margins. But in terms of actual performance, I'm not holding my breath until Apple can prove to me they can design an ARM chip that outperforms a Core i5 at the very least.

And I'm pretty sure someone in here will cite Geekbench, but Geekbench can hardly be qualified as a "complex" workload.

That article is from 2014. It's 2020. Intel is still stuck in 2014 and Apple has chips in a $400 phone that are almost as good as what you get in a $2,000 laptop (Windows or Mac).

I'm not worried about the raw horsepower of an ARM Mac, it will be great. I'm a lot more worried about the experience using the programs, and how well optimized third party vendors make their software for it. MacOS has gotten extremely buggy, and they really need to focus on stability not features.