Mac Transition to Apple Chips

[R C-R]

6 minutes ago, v_kyr said:

That's what we already said before in this and the other related threads, so you are just repeating things here that we all already know and had discussed long before.

I was responding to your "Well the CPU doesn't play any role here..." comment, which seemed to me at odds with what has already been discussed. What the software can do depends strongly on what the hardware can support, which surely includes the CPU, right?

Besides, the CPU is not the only hardware that processes data. Many GPUs can run many processes in parallel hundreds of times faster than even the most powerful CPUs. Some processing tasks can be run entirely in T series coprocessors. Processing is less centralized into one 'unit' than ever before, so doesn't it follow that we have to consider the contribution of all the Apple chips (collectively "Apple Silicon") here?

[Stephen B]

I would really like to hear from Affinity how they plan to do it.

This is going to my third major hardware architecture transition on the Mac and especially the PPC to Intel one was horrid - everything became so bloated, slow and buggy that we used one dedicated machine for testing and didn't touch production computers for almost two years. By then it was a clean cut.

Point in question: are Affinity planning to release dual apps (one app runs natively on either hardware) or will there be two versions, i.e. one native Intel and one native ARM app with the same functionality? If so, how long do you plan to support the Intel Macs with identical features and bug fixes, version per version?

Depending on the answer I might update some machines to the latest Intels and try to weather out the transition period.

Fact is, each of these transitions were a nightmare for many software publishers. The older and/or complex their code base was, the more difficult the transition became. I really don't envy the developers right now, and maybe for the next two or three years.

[v_kyr]

36 minutes ago, R C-R said:

I was responding to your "Well the CPU doesn't play any role here..." comment ... and before ... but somehow I don't think the interfaces of Apple TV's, Watches, or iPads will ever look or function the same way as they would on Macs, even if future versions of all of them are running on Apple Silicon.

The interfaces and UI stuff etc. are pretty CPU agnostic here, nowadays CPUs & (integrated) GPUs are all by far powerfull enough to show them up, no matter if a X86/64 or RISC based main CPU is used. They could every time change all that and let it run on other CPU/GPU chips too whenever they want to.

49 minutes ago, R C-R said:

Besides, the CPU is not the only hardware that processes data. Many GPUs can run many processes in parallel hundreds of times faster than even the most powerful CPUs. Some processing tasks can be run entirely in T series coprocessors. Processing is less centralized into one 'unit' than ever before, so doesn't it follow that we have to consider the contribution of all the Apple chips (collectively "Apple Silicon") here?

What does that have to do with their user interface changes like app groups, rounded rects, more skeuomorphistic icon looks etc. here, those aren't Apple Silicon dependent at all.

[R C-R]

16 minutes ago, v_kyr said:

What does that have to do with their user interface changes like app groups, rounded rects, more skeuomorphistic icon looks etc. here, those aren't Apple Silicon dependent at all.

None of us know what interface changes Apple will make in the next few years, but I suspect more than a few of them will be dependent on things that Apple won't support on Macs running on less energy efficient CISC chips.

It would not be the first time Apple has done such things. It also seems consistent with what Tim Cook said about embracing big changes to make the Mac "stronger and more capable than ever."

Of course, this means a lot of mac users will be motivated to buy new Apple silicon powered Macs, but I do not think Cook will see that as a problem. 🙂

[wonderings]

1 hour ago, Stephen B said:

I would really like to hear from Affinity how they plan to do it.

This is going to my third major hardware architecture transition on the Mac and especially the PPC to Intel one was horrid - everything became so bloated, slow and buggy that we used one dedicated machine for testing and didn't touch production computers for almost two years. By then it was a clean cut.

Point in question: are Affinity planning to release dual apps (one app runs natively on either hardware) or will there be two versions, i.e. one native Intel and one native ARM app with the same functionality? If so, how long do you plan to support the Intel Macs with identical features and bug fixes, version per version?

Depending on the answer I might update some machines to the latest Intels and try to weather out the transition period.

Fact is, each of these transitions were a nightmare for many software publishers. The older and/or complex their code base was, the more difficult the transition became. I really don't envy the developers right now, and maybe for the next two or three years.

They showcased Rosetta 2 at the release. This is the software that helped with the transition from PPC to Intel. I do not remember having much headaches as the big apps all moved over pretty quickly. If memory serves me I switched from our Dual G4's and Dual G5's to intel Macs early on which would obviously make the software transition easier not needing to be running on legacy cpu's. I am sure there was some stuff I grumbled about back then, but nothing specific comes to mind. 

At WWDC Apple showcased a reasonably high end video game playing on an Apple running their own CPU. The app was made for Intel and was being translated by Rosetta 2. The demo looked great, colours and movement were fluid in a graphically rich 3D environment. Of course who knows what was fine tuned on the back end. That being said we have come a long way since the PPC ships and Apple has been using their own processors in their phones and tablets successfully. I like to think the transition will be a smooth one. Of course there will be some issues and those I believe will generally come from products no longer being upgraded or apps from smaller developers who do not have the resources to upgrade or are just not working on it anymore and thus the software will eventually be lost to anything non intel. 

[v_kyr]

• Port your Mac app to Apple Silicon

[JGD]

On 6/28/2020 at 6:56 PM, v_kyr said:

Yes it was initially developed on other CPUs than the one used by black hardware (Motorolla CPU). Later the NS/OS was adapted/ported over to other CPU platforms, among those the Intel x86 CPUs (used by IBM & Clone PCs aka Dos/Win machines those days) and Sun and HP RISC Workstations. Development and porting over of apps etc. was pretty easy and almost just a project recompile with enabled settings for code generation for all four architectures, which resulted into fat binary (universal binaries) executables.

I think the main point for Apple is here instead, to be independent from other chip manufactors (Intel) and so to have the whole control of chip manufactioring, it's capabilities and availability in their own hands. And as one can see from the chips they used in former times (Newton devices) and nowadays iPhones/iPads or their T2 and other custom chips, they are long time able to do so. They have long time experiences with the ARM CPU architecture and also the tools to make use out of those equally good.

Sure, that was never in contention. I'm just saying that “hackers gonna hack” and, if the demand is there and the returns justify it, they will cave in to the market pressure. When they did the switch to Intel, Windows compatibility was never their main goal (if at all, which I also think it wasn't, as they ditched the vintage BIOS spec but never adhered to UEFI when it did become the norm on the Wintel world) and, yet, since that compatibility was there ripe for the taking and the investment on drivers for such a lean line-up was negligible, they went ahead and created Boot Camp. And the rest is, as they say, history.

Them coming up with a Boot Camp 2, on an ARM-dominated or at least ARM-infested PC world (to the point that universal Windows x86-64/ARM-64 binaries also become the norm, that is), wouldn't be totally farfetched. Their SoCs would still likely eat Qualcomm's for lunch and make Macs, more than ever, the best PCs around for running any OS, so it would be a win-win situation, if you ask me.

Then again, there's the whole virtualization angle. Perhaps they have an ace up their sleeve with that one? Maybe some great new way of doing it that actually makes the overhead of macOS also negligible? The very fact that they emphasized virtualization so much on the WWDC keynote, and not on some random session, is a dead giveaway that they must have some grander plans for it in the future. With Apple, more often than not, you have to do a lot of “Kremlinology” and tea-leaf reading; it's usually what's left unsaid that is truly relevant. Surely they don't expect Linux to be the only alternative OS to ever run on Parallels/VMware/Virtual Box (by the way, there was no mention of macOS itself running on a VM, but it also damn well should, especially considering how Rosetta 2 will eventually be axed, just like Rosetta [1, the PowerPC variant] and Classic/Blue Box were before it)…

[v_kyr]

Hard to tell at the moment, since they didn't told in WWDC sessions much about possible virtualization support, or the future of tools like their Boot Camp etc. at all. - Instead they talked more about how to port Intel CPU based mac apps to the upcoming ARM64 architecture and what steps have to be taken into account to successful manage these tasks. So most related things here will have to be discovered, have to be found out and shown over time. At the moment it is still difficult to make specific and concrete forecasts here.

• Port your Mac app to Apple Silicon

• Explore the new system architecture of Apple Silicon Macs

However, for apps like the Affinity line, it will probably take some time until those will be available for the ARM architecture then, since every needed code dependency (all by Affinity used third party libs etc.) will also have to be first natively compiled and ported over here. Thus all that the AF iPad versions still don't offer in contrast to macOS versions yet, will need to be adapted/ported over then.

[McD]

On 6/27/2020 at 11:42 PM, JGD said:

Well, I see where you're coming from with that. But let's face it, not even System/Mac OS 1.0 was entirely original, it was a wholesale copy of Xerox's Star with a lick of Bill Atkinson, Andy Hertzfeld, Steve Capps, Susan Kare et al.'s paint on top of it. And Xerox Star was a platform which, in turn and mind you, probably wouldn't even exist – or at least definitely not in that particular form – if it wasn't for Douglas Engelbart's “mother of all demos”.

As for more recent times… could you seriously live productively without Notification Centre on iOS/iPadOS? That's also an Android innovation right there. Resizing from all corners and sizes on macOS? A belated addition to then-called OS X, lifted wholesale from Windows. And the same goes for the current fullscreen and split screen implementation… As for Maps' features? Well, since Apple ditched Google as a default provider many years ago and, has Marques Brownlee keenly pointed out, still don't allow you to set a different mapping app as the default, they'd better copy them ASAP globally (I live in a small country in Europe where Apple has always lagged behind, and it kind of pisses me off that the world's biggest company still hasn't properly caught up with Google)…

And the iPod? Was it the first pocket, mass storage-based music player? Was Apple Watch the first modern smartwatch? Nope, they were just the best at their game from day one. Time and time again you see Apple playing out that MO: occasionally they come up with an explosive innovation which quickly propagates across the entire industry (say, the PowerBook 100, which was one of the first such products on the market with the pointing device closer to the user and the keyboard farther, the PowerBook 500 and its seminal trackpad, multi-touch gestures on the 2004 PowerBook and iBook G4 trackpads, or the entire touchscreen phone and tablet paradigm… the latter of which, as you very well know, also originated and were demoed outside of Apple well in advance of their conversion into workable products: https://www.youtube.com/watch?v=89sz8ExZndc) or ballsy, pioneering move (ditching the floppy, serial, parallel and SCSI ports in one go, then the optical drive, then their own FireWire and shortly after USB-A and all other I/O), but they mostly copy others with a twist or a flair. Steve Jobs himself put it out on display, shamelessly, with his apocryphal quote “Good artists copy, great artists steal”: https://quoteinvestigator.com/2013/03/06/artists-steal/

I don't mind copying in principle. Everyone does it, and I stand by it even on creative fields like my own (remix culture, as defined by Lawrence Lessig, comes immediately to mind) and am deeply skeptical of the current climate of copyrighted *everything*. Heck, I work in type design, an area where we've been copying Roman artisans and Irish scribes for around 2000+ and 700+ years respectively. I'm fine with it, as long as a modicum of ethics is maintained in said copying – namely, by giving some form of credit and/or compensation – and some common advance of the human species is achieved – that should go without saying… If you're bringing nothing new to the table, you are indeed only taking advantage of others for immediate personal gain. I won't make any excuses for Apple when they weren't on the right side of history (can you say “Sherlocking”? Konfabulator? Growl?), but other than those few examples they did get better at it once they had the resources to outright acquire good ideas and properly compensate their originators whenever possible. For every of those bad moments, there are many more when they did things right (for instance, SoundJam MP, Coverflow, Beats, Workflow or Shazam come to mind). As for two behemoths (historically Microsoft and Apple, now mostly Google and Apple and sometimes Samsung and Apple) constantly copying one another? I couldn't care less, as long as end users collectively benefit from it.

Wow. Someone really can’t differentiate between a component & a product.

I particularly liked the part where iPhone notifications were copied from Android as opposed to macOS or Growl. Most ‘Android innovations’ were straight lifts from previous desktop OS’

[McD]

On 6/29/2020 at 10:52 PM, v_kyr said:

In regards to the WWDC announcements, namely?

Other than that many people have different feelings for the software UI changes ...

• Hey designers I made a tutorial for migrating your icons to Big Sur

 

GPU & Neural ISAs (which do most of the work) are Apple’s 1st party ISAs.

[Dan C]

On 6/22/2020 at 9:05 PM, puredesign said:

Will the Affinity apps be ready for the transition to Big Sur and the Apple chips when new hardware ships?

Apologies for the delayed response here!

We absolutely do plan to support Apple Silicon from the day it becomes available & our apps will be compatible with the official release build of macOS Big Sur running on either Intel or Apple Silicon hardware   

[Brad Brighton]

On 6/29/2020 at 6:24 AM, Stephen B said:

I would really like to hear from Affinity how they plan to do it.

This is going to my third major hardware architecture transition on the Mac and especially the PPC to Intel one was horrid - everything became so bloated, slow and buggy that we used one dedicated machine for testing and didn't touch production computers for almost two years. By then it was a clean cut.

Point in question: are Affinity planning to release dual apps (one app runs natively on either hardware) or will there be two versions, i.e. one native Intel and one native ARM app with the same functionality? If so, how long do you plan to support the Intel Macs with identical features and bug fixes, version per version?

Depending on the answer I might update some machines to the latest Intels and try to weather out the transition period.

Fact is, each of these transitions were a nightmare for many software publishers. The older and/or complex their code base was, the more difficult the transition became. I really don't envy the developers right now, and maybe for the next two or three years.

Short version: The vast majority of the style of cruft that you're referring to that made PPC->x86 painful was/is weeded out by the 64bit-only transition well before the Apple Silicon transition. If a developer wants to bring something pre-64bit over, yes, it's going to involve work.

If that work has already been done or if the app is natively friendly to the current world anyway, this transition promises to be _relatively_ painless (excepting certain categories like device and/or kernel drivers, etc) which is more a function of the increased security layers that the OS is getting than the drive to Apple Silicon itself.

[JGD]

On 7/6/2020 at 12:46 PM, McD said:

Wow. Someone really can’t differentiate between a component & a product.

I particularly liked the part where iPhone notifications were copied from Android as opposed to macOS or Growl. Most ‘Android innovations’ were straight lifts from previous desktop OS’

What difference does it make whether it's a component or a product we're talking about? That's not the crux of the matter…

Many of the examples I've mentioned were, in fact, bolted-on, third-party components, even if they were sold or otherwise distributed as products, and which became, one way – through acquisition – or another – through wholesale rip-off –, first-party ones. And the same happened with whole applications, and even entire GUI philosophies. I don't get why you're so hung up on the scale of the copying itself, as what really mattered to me in that post was the way it was done and the size of the parties involved themselves.

The TL;DR of it is: it only irks me when the 800 lb. gorilla rips off the tiny indy dev; I couldn't care less when the 800 lb. gorilla rips off another 800 lb. gorilla. Especially if it's something obvious, essential and not especially creative like notifications or widgets (it's not like those are “aha!” innovations like, say, multi-touch as an entire interaction model), which ends up benefiting large numbers of users.

[JGD]

On 7/2/2020 at 2:36 AM, v_kyr said:

Hard to tell at the moment, since they didn't told in WWDC sessions much about possible virtualization support, or the future of tools like their Boot Camp etc. at all. - Instead they talked more about how to port Intel CPU based mac apps to the upcoming ARM64 architecture and what steps have to be taken into account to successful manage these tasks. So most related things here will have to be discovered, have to be found out and shown over time. At the moment it is still difficult to make specific and concrete forecasts here.

• Port your Mac app to Apple Silicon

• Explore the new system architecture of Apple Silicon Macs

However, for apps like the Affinity line, it will probably take some time until those will be available for the ARM architecture then, since every needed code dependency (all by Affinity used third party libs etc.) will also have to be first natively compiled and ported over here. Thus all that the AF iPad versions still don't offer in contrast to macOS versions yet, will need to be adapted/ported over then.

Thank you for the links!

Well, let's hope they get around to solving those dependencies and recompiling their code soon. Everything points to that, and I'd say their pervasive use of platform-agnostic (if not completely ISA-agnostic) C, as well as Metal (which, I'm guessing, in the context of 2D apps shouldn't be so complex as to require much rewriting), should help on that… From all I've read here on the forums before, Serif always seemed to keep a good balance between keeping things abstracted enough from the hardware, while at the same time making use of native APIs and frameworks.

I don't know how they manage it, and I still feel their apps could and should look and feel more native, but they're still leagues better than Adobe apps on that regard (well, if anything, they don't have to deal with decades-old spaghetti code  ). I'm willing to bet that they'll transition their current three Mac apps way quicker than Adobe will (the absence of universal Illustrator and InDesign demos was notably conspicuous, for one).

And when I say quicker, I say “way before the Big Sur GM even drops”, which might be earlier than the first Apple silicon Mac release. Of course, we won't know anything about that, as it'll likely be heavily protected by Apple's NDA, but we'd be naïve to think that Serif doesn't have an Apple DTK being shipped on their way or already set up on their offices somewhere. That's the only way they can promise to have it up and running on day one.

[Andy Somerfield]

Morning all,

Further to the comments made by Dan, I'm pleased to report that the transition to Apple Silicon has been completed. We will, of course, use the time between now and any Apple hardware release to test what we have internally - but I see no reason why we will not be available on day one with all three apps.

To address questions regarding ongoing support for Intel based Macs: We will support Intel based Macs as long as they are supported by Apple (ie. until the last Intel based Mac is considered "obsolete" by Apple). This will be years after the last Intel Mac has shipped - and will only mean that we stop providing updates - not that the app will stop working.

Hope this helps,

Andy.

[ronnyb]

43 minutes ago, Andy Somerfield said:

Morning all,

Further to the comments made by Dan, I'm pleased to report that the transition to Apple Silicon has been completed. We will, of course, use the time between now and any Apple hardware release to test what we have internally - but I see no reason why we will not be available on day one with all three apps.

To address questions regarding ongoing support for Intel based Macs: We will support Intel based Macs as long as they are supported by Apple (ie. until the last Intel based Mac is considered "obsolete" by Apple). This will be years after the last Intel Mac has shipped - and will only mean that we stop providing updates - not that the app will stop working.

Hope this helps,

Andy.

That was quick!

[AffinityMakesMeSad]

4 hours ago, Andy Somerfield said:

Morning all,

Further to the comments made by Dan, I'm pleased to report that the transition to Apple Silicon has been completed. We will, of course, use the time between now and any Apple hardware release to test what we have internally - but I see no reason why we will not be available on day one with all three apps.

To address questions regarding ongoing support for Intel based Macs: We will support Intel based Macs as long as they are supported by Apple (ie. until the last Intel based Mac is considered "obsolete" by Apple). This will be years after the last Intel Mac has shipped - and will only mean that we stop providing updates - not that the app will stop working.

Hope this helps,

Andy.

Lovely!

If we remember right from the WWDC2020 Apple said that big advanced apps on Mac can be up and running in two to three weeks, so, Serif is working hard here!

And, the transition to other apps could be just days or hours to run 100% on Apple Silicon...

[Dave Tuck]

Thanks Andy that's very reassuring.  I'm heading over to Apple to order my new machine.

😀

 

[JGD]

On 7/9/2020 at 6:59 AM, Andy Somerfield said:

Morning all,

Further to the comments made by Dan, I'm pleased to report that the transition to Apple Silicon has been completed. We will, of course, use the time between now and any Apple hardware release to test what we have internally - but I see no reason why we will not be available on day one with all three apps.

To address questions regarding ongoing support for Intel based Macs: We will support Intel based Macs as long as they are supported by Apple (ie. until the last Intel based Mac is considered "obsolete" by Apple). This will be years after the last Intel Mac has shipped - and will only mean that we stop providing updates - not that the app will stop working.

Hope this helps,

Andy.

As I suspected, that was quick! It's good to know, and further confirms my theory on why Apple would rather use Affinity Photo as a demo for Rosetta 2 (which, mind you, feels a bit like cheating, because Affinity apps are lean enough to make them somewhat usable even under binary translation), rather than as a poster child for Universal 2.

Adobe users are still stung by Quark having dragged their feet during the transition from Classic to then Mac OS X, which led them to switch to InDesign in the first place, and by Adobe, in turn, having dragged their feet during the transitions from Carbon to Cocoa, from PowerPC to Intel and from macOS to i[Pad]OS. Seeing how those users are still in the majority, it would be suicidal of Apple not to prioritize those apps and pretty much ensure they would be universal on day one…

As for Serif apps, those were among the first desktop-first apps to appear on the iPad. It was patently obvious they would be the first to run natively on Apple Silicon macOS, even without advance warning from Apple. Adobe, on the other hand, had to be given early access to the DTK and likely extra help from Apple engineers. It's both sad and impressive to see the difference in treatment here…

[NewUserr246]

Hello everybody! Does anybody know if Affinity photo will run on a DTK presently? I would like to purchase one for my dev work if it can indeed run on a DTK. Or should I wait for Apple's Silicon hardware release before purchasing Affinity? Thank you! 

Edited September 14, 2020 by NewUserr246

[Blende21]

Maybe you should read the thread you are posting in, before asking an answered question again.