“This week Apple plans to introduce new Macs to an aging lineup and the invitation to the October 27th event was a simple ‘Hello Again’ card; white text, black background, partially covered Apple logo,” Wil Gomez writes for Mac360. “The original Mac was introduced with a simple ‘hello’ written in cursive, way back in 1984. ‘Hello Again?’ It’s new Macs. But maybe something else. Perhaps a Mac with it’s own CPU inside instead of chips from Intel.”
“Let’s think thinner, faster, lighter for the MacBook Pro plus some kind of Magic Toolbar with Touch ID built-in, perhaps a speed bump for the iMac, maybe even some USB-C ports in the old MacBook Air and a lower price tag,” Gomez writes. “What else could come from Apple’s ‘Hello Again‘ show? ARM-based Apple A10 Fusion-like CPUs in the MacBook Air and the Mac mini.”
Gomez writes, “This would be a great way to introduce a far less expensive Mac line without devaluing the the value of the more capable MacBook, MacBook Pro, iMac, or Mac Pro, all of which can be made faster with new Intel CPUs inside.”
MacDailyNews Take: As we wrote just after Apple’s invitation was delivered, “The use of ‘hello’ by Apple (previously used to introduce the original Mac, the original iMac, and the iPhone) is significant and signals that this is more than a typical Mac event.”
Now, does it make more sense to be smearing your fingers around on your notebook’s screen or on a spacious trackpad (built-in or on your desk) that’s designed specifically and solely to be touched? Apple thinks things through much more than do other companies. The iPhone’s and iPad’s screens have to be touched; that’s all they has available. A MacBook’s screen doesn’t not have to be touched in order to offer Multi-Touch. There is a better way: Apple’s way. And, no Gorilla Arm, either.
The only computers using Multi-Touch properly, using device-appropriate Multi-Touch input areas are Macintosh personal computers from Apple that run OS X (and Linux and can even slum it with Windows, if need be) and iOS even more personal computers (EMPCs), namely: iPhone, iPod touch, iPad, and iPad mini.
Note that none of this bars a “MacPad” from production. Any iOS-based iPad would become a high quality display (possibly still “touchable,” but likely not due to the reasoning stated above) when docked into a “MacBook” (running OS X, and providing keyboard, trackpad, processor, etcetera). Such a convertible device would negate having to carry both an iPad (car) and a MacBook (truck) around. They’d be one thing, but able to be separated into two, each providing the best capabilities of their respective form factors. — MacDailyNews, May 4, 2013
—
Think code convergence (more so than today) with UI modifications per device. A unified underlying codebase for Intel, Apple A-series, and, in Apple’s labs, likely other chips, too (just in case). This would allow for a single App Store for Mac, iPhone, and iPad users that features a mix of apps: Some that are touch-only, some that are Mac-only, and some that are universal (can run on both traditional notebooks and desktops as well as on multi-touch computers like iPhone, iPad, iPod touch, and – pretty please, Apple – Apple TV). Don’t be surprised to see Apple A-series-powered Macs, either. — MacDailyNews Take, January 9, 2014
—
Anyone in the market for a 12.9-inch device that’s an OS X-powered MacBook when docked with its keyboard base and an iOS-powered iPad when undocked?
Illustration from Apple’s hybrid Mac-iPad patent application
I think Apple is going to announce products that will not be immediately available. I think it will introduce new MacBook and MacBook Pro but I also hope the announce something along the line of an iMacPro or MacPro. I am currently in the market for said machine. The MacPro in it’s current form does not represent good value and Apple knows it.
“The MacPro in it’s current form does not represent good value and Apple knows it.”
Boy is THAT an understatement. In my opinion there should never be a moment when that situation should exist. Even though I know that’s easy for me to say.
I think there will be some updates to their desktop lines. They almost have to at this point. I doubt they will spend much time on it. They will probably show off updated FCPX too. This will most likely be at the beginning of the show and they will quickly move on to the MacBooks. That will be the highlight of the show, and the only thing the press will talk about. Watch the show, re watch it, then visit Apple’s web page to get any information on desktops. I am sure Tim will squeeze in IBM’s new TCO study at the top. Probably at the expense of desktop time on stage.
I want a new Mac Pro as well (an UPGRADABLE machine no less)
looking at the issue at another angle: the MP is built in USA, can Apple afford to abandon building Macs in USA? the political pressure to continue is great as Tim Cook keeps having to defend his position of jobs in USA (latest hits is during the presidential campaigns). In a TV interview a few months ago when Cook was asked about USA manufacturing he NEVER mentioned the MP, probably because he knew it was so out of date, which probably galled him. So here’s just hoping for political reasons (all those ‘ban’ made in China protests etc) T.C will keep MP USA alive (but NOT the Cylinder please, some UPGRADABLE TOWER would be better…)
I grant that the cylinder was not an ideal shape. But had they chosen to they could have made it upgradable using a stacking configuration. The cylindrical cards and housings would have been more expensive but it could have been done and Mac Pro users would have gladly paid the difference. I love Apple but sometimes I just can’t fathom what they’re thinking.
The A10 variant for Macs (A10M) could include 8 or more cores at higher clock speeds (compromise on power efficiency and heat production) to increase performance. And Apple’s multi-core GPU embedded in the A10X already kicks 4K/SUHD butt as-is. But Apple could add some more GPU cores, too.
Who knows where this might go. Your skepticism is premature.
Perhaps. Architecture changes need a WWDC.. Because other then all the little app store apps all the big boys would need to re-code for yet another apple processor flavour of the year. Or.. just say F it and keep releasing for workstation class powerhouses running linux and windows.
agree, it wouldn’t make sense to announce an architecture change without giving the devs lead time to make changes. they at least give the summer to update from iOS9 to 10 and from 10.11 to 10.12
That was my thoughts as well….I hope not. Or if they do have a separate line of them along side the intel versions. Moving away from intel and loosing Windows compatibility I think is a mistake if true.
Traditionally, the “Hello” thing has accompanied significant changes of direction rather than incremental upgrades. It seems quite plausible that Apple could be about to announce A-series CPUs in Macs.
I’m convinced that the A series Macs will be a range that sits alongside Intel Macs. Those who need Intel Macs will pay the extra for the Intel CPU, while those who use only applications written for Macs will be able to enjoy cheaper and more power-efficient Macs using A-series CPUs.
I’ve never needed Windows compatibility and would buy an ARM powered Mac as soon as the applications that I use are available for ARM.
The problem with dual architecture options is that Apple has been overrun by beancounters and can’t seem to evaluate each product on its own merits, only by percentage of overall revenue.
Look at the Mac Pro. They generate the least total revenue in a quarterly report even though per-machine margins are high, because it’s a niche market. And even though that niche is important, Apple has ignored it for 3 years (and arguably longer, if you consider that the Mac Pro form factor is unsuited for proper pro work in many ways).
Even Macs in general have suffered, with no major updates to most families in over a year. If we have A-processors in the least expensive offerings, which traditionally are what sells the most, the beancounters can say “hey look, the Intel line isn’t selling well” and convince the execs (if the execs aren’t the very shortsighted beancounters in the first place) that not enough customers are interested in Intel compatibility, leading to either higher-end Macs languishing, or cut altogether.
I’m convinced it is going to be a big announement. THey are obviously channeling other historic announcements. Add that to the time the Mac has languished plus the effort Apple is known to put into their products. I’m expecting a lot.
That was my thoughts as well….I hope not. Or if they do then have a separate line of them along side the intel versions. Moving away from intel and losing Windows compatibility I think is a mistake if true. Also the amount of time between upgrades is another sign of major changes coming.
iPhone and iPad with A10X. Able to plug into the new Thunderbolt Display. 256Gb storage option on these devices allows them to become the primary computing device.
Walk around with your primary and only computer in your pocket. Full size apps when you plug into the screen at home but work as normal iOS variants when screen unplugged.
Switching processors brings up the issue of getting old code to run on new processors. How does Apple pull off a processor switch once again? Is there a version of Rosetta that translates x86 to ARM instructions?
Apple does not need to use the Rosetta approach again. Apple probably already has Xcode ready to compile for ARM. Same source code – multiple executable targets.
No actually. You still don’t understand the difference writing for CISC CPUs and RISC CPUs. It is far from as simple as you pretend. I keep pointing the way and if you don’t go there you end up spreading nonsense about this situation.
As a software engineer and previously a RISC processor ALU designer I disagree.
Yes, there will always be complications through any transition, but I disagree that this is an impediment or would stop Apple from doing whatever is most compatible with where they see computing in 5-10 years.
A focus on the product results in growth for Apple down the road, even as it often generates pain for users in the short run.
On the technical side, most code can just be recompiled. Not all, but most as in 99.5%. Of course the 0.5% that must be rewritten is often very foundational/complex low-level processor/device specific code. But it is in the minority.
I would see it as Apple’s job to provide as much low-level transition support as possible (BLAS libraries, make real gurus available for developers, etc.) while making most code just recompile seamlessly.
Apple has the advantage of owning their whole tool, software, OS and hardware stack, and even an optional software store for handling sales, updates, etc. They won’t move until they think everything is lined up right.
And I agree with the Intel hold outs, Apple will continue shipping Intel Macs. It will be the low end that gets A-series chips first where performance and Windows compatibility are not issues for most people.
I profoundly disagree about the required code rewrite. Don’t foist nonsense please. People tend to believe it these days. The usual advice: Study CISC vs RISC. You clearly have not.
If you actually read what he wrote, he agreed that most of the code could simply be recompiled. Rather, a small percentage of code would have to be rewritten. As someone who has ported applications across architectures (same OS), I agree with his assessment and would suggest that arguing this point simply demonstrates that you don’t know what you’re talking about. Generically speaking there could be endian issues to deal with, but even that isn’t likely. Your CISC vs RISC comments are pretty comical. If you’re writing in Swift, how exactly do you expect to be impacted by this? If you’re writing in assembler, it would be an issue, but then you’d need a complete re-write anyway.
Go read the reply I wrote to ‘Michael’. I am not going to laugh at you for your being wrong, so don’t worry.
To be clear, I’m not agreeing with KingMel’s position that something like Rosetta wouldn’t be needed. It would. Rather, I’m laughing at your explanation as to why it would be needed. Perhaps you might want to rethink your reasoning before commenting further. Hint: CISC vs RISC really has nothing to do with it.
Thanks for making factual arguments. I agree with them.
Yes, porting low-level code will take work. Apple will no doubt take that into account and work to make that as easy as possible. Their may be some emulation, but certainly tools to help with that.
But don’t expect Apple to hold off from what it sees as a long term advantage (if it does see owning its Mac processors as advantageous) due to the short term pain for a minority of developers.
Product comes first for Apple, not developers. They have truly screwed developers many times. Developer pain is not hight on the priority list for for Apple as they have demonstrated many times.
Yes, I agree with your position that Apple will do what Apple thinks is best, regardless of developers needs, etc. I don’t believe there is much of a technical barrier for Apple at this point. There are other things to consider though. Chip development is expensive. Apple certainly “can” build a better chip than Intel based on ARM’s instruction set. I don’t think that’s in question. The question is about benefits.
It makes sense for Apple to do custom silicon for their iPhones because the iPhone is Apple’s “bread and butter” and because Apple ships products in ridiculously high volumes with these chips. Especially when you include iPads, iPod Touch, AppleTV, etc. The volume of Mac shipments is very low by comparison and would hardly justify the cost of a custom CPU, etc. Further, the ability to run Windows apps (boot camp or virtual PC) has proven to be an overall benefit for Apple in selling Macs. It’s not clear that creating a custom Mac ARM based CPU would be worth it.
Derek, I would rarely disagree and would almost never be read to you.
But f#ck you. I have designed chips. I know the difference between RISC and CISC far better than you probably do as have designed instruction sets for both.
Your faceless ad hominem attack is not like you, so I look forward to more sensible discussions in the future.
1) RISC to CISC instruction set conversion is easy. Their are profound differences in how they pipeline and branch predict, etc., for performance but the actual translation for equivalent functional behavior is EASY when conversion is from C or another higher language in both cases.
2) x86 is not just CISC but CISC laden with enormous design debt. It is actually very inefficient in terms of silicon area compared to what the same CISC functionality would have if designed in modern times.
3) RISC by definition has very little design baggage as there is so much less to complicate.
Get a f@cking education on RISC and CISC – by which I mean actually go through the process of designing instruction sets and fabbing processors. I HAVE.
And when you communicate a view BACK IT UP with objective points. I continually see you making CISC > RISC comments without the least bit of factual information.
You are wrong to start with. But when you don’t back up your points objectively you are not even wrong.
Derek, I apologize for coming on strong, but not for strongly disagreeing with you from real experience.
Normally I agree with just about every view you express.
As I posted elsewhere today, I’m going to sit back and snooze through this nonsensical subject into the future. Have fun watching nothing happen, for the reasons I’ve been jabbering for a year now, and more… 💤💤
I am not looking forward to transitioning to a new cpu architecture again. And, I don’t think that is what is going on. I think it more likely that we will see a new device. One that is between the ipad and a macbook. I believe MDN has said that these two platforms will continue to merge/blend over time.
Obviously yes. Thank you for pointing that out. Moving the Mac to A-Series RISC CPUs would cause outrage and uproar among ALL Mac developers. That’s one of the several reasons this is a Nonsense Meme.
Nobody complained about developing for ARM/iOS/iPhone/iPad, or for watchOS. As long a there is a compelling advantage for customers to have A-series laptops (say cheaper, faster for the money, longer batter life, etc) developers will be fine. I am one of them.
What would be bad would be if Apple announced A-series laptops without a big customer-centric reason for getting them. That would be a disaster for Apple and developers would just ignore the new machines.
Apple knows this. They are not just going to swap CPUs. If they swap CPUs it will be to deliver something obviously better for a well defined and large group of customers.
Oh, look. I get to repeat myself. Sorry, too bored to bother.
I seriously doubt there will be a switch to ARM for Macs. There is a lot of reasons not to it, with few reasons to do it. This has been in the press for years, like the Apple HDTV that everyone knew was going to come “next” year.
Like most Apple events the press will report on some wild ideas they think Apple will do, then complain that Apple did not do what they made up. They will spend little space on what the new Apple product can do because that would mean real journalism. The ARM Mac is another shot in the dark so if it does happen then people will say I right, see how smart I am.
Wish Apple would just buy an x86 license and take their CPU knowledge and make their own chips for the mac. If they did that, they wouldn’t need to shift every mac customer/developer to a new architecture, and they could incorporate all they know about their A series CPUs in the new x86 chip.
If Jobs were still running Apple, I think we would have already seen the switch away from Intel. There is really no major reason to stay with Intel anymore, when they have their own chip design. The performance is clearly there, the ability to ramp up manufacturing as well, their chip design team is extremely successful at making major strides each year.
Let us not forget; the move to Intel wasn’t so that the Mac could run Windows (although at the time, with significantly lower market share, as well as strong marketing push, it was a bonus). It was because IBM/Motorola weren’t delivering. Today, Intel’s roadmap isn’t any more ambitious than Apple’s own for CPUs. Bringing everything under their own control reinforces the philosophy of designing your own software and hardware.
There are very few Mac users who still run Windows. Their numbers are negligible; probably fewer than those who still use optical disc on a Mac. Apple has already eliminated that optical disc (as they did many other features and ports).
Again, if Jobs were still alive, he would have already moved over, and pushed everyone (Adobe, Microsoft, other big players) towards the new architecture. However, without Jobs, I’m not so sure.
I’m not so sure. Were he alive today, Steve Jobs would have changed along with shifting market preferences. His predilection for guillotining established tech standards in favour of proprietary new-age gimcracks would have by now matured into a real appreciation for the angry pain of transition suffered by so vast a number of sorrowful pro users. His humanity had increased to the point of actual compassion by the time he passed. Rather than drag everyone screaming into the future, as he once tried to do, he’d have learnt not to blow off a hat-in-hand emailer with an abrupt “you’re holding it wrong,” but to turn around and yell at his engineers that “you’re designing it wrong.”
Please cite your source for the statistics about the vanishing base of Mac users who run Windows. It seems too careless an assumption, and not consistent with my experience. You once pointedly cited studies of gunshot death rates by country to argue for the efficacy of gun control. Is not the pattern of OS usage as worthy of statistical support? Granted, nobody running Windows actually dies. They only wish they had… 😬
The days of intel consumer computers is nigh, so it may not be this year but far sooner than people think. Old technology that is kept around for the some purpose of lack of vision and it being today’s bread and butter product pushed by the sales staff, is ALWAYS the death knell of EVERY company that has ever existed.
Intel doesn’t seem to be pivoting the same way that IBM or Microsoft has. It looks like it will be going the way of Smith Corona.
The only way Apple would move to ARM based processors for the Mac line would be either a) to significantly reduce the hardware build cost of the Mac, so they could enter a lower price tier without giving up margins, or b) they want to migrate everything to IOS, and that would be a temporary means to start moving in that direction. They could choose either reason, but I do not believe they will.
I suggested years ago that Apple build an IOS compatibility module into MacOS so that IOS Apps could run on both platforms, but nobody listened back then.
Sources tell me also iMac may not even get refreshed in this announcement.
I wonder whether the ‘space’ like design of the invite shows an upgrading of cloud infrastructure, maybe some kind of A powered Macs to compete with Chromebooks, especially for areas like education where Apple marketshare has dropped to about 10% and Chromebooks climbed to 50% (the education market is a very important area for Apple as besides revenue it trains new users).
An iCloudBook would be a good idea for education, if Apple were able to hit near Chromebook pricing. There is also a segment of consumers whose computer activities are already Internet-based, meaning they use a web browser, email and social media. So requiring an active internet connection wouldn’t be a huge change from their use pattern. That would be interesting for the low-end, if Apple decided to play in that space. Typically Apple has fancied itself a Mercedes and not a Kia.
No it doesn’t. Not gonna happen. I’ve been over the details of why this is a nonsense meme at least 10 times and I won’t be going over it again. Go look up ‘CISC’ and ‘RISC’ and learn why for yourself. Just don’t bother me with your lack of comprehension. Thank you.
You do know that Intel’s CPUs are in fact RISC cores and have been for many, many years? “CISC” instructions are broken down (transcoded) before being passed along to the actual RISC core to be processed.
Starting with Pentium Pro (P6 microarchitecture), Intel redesigned it’s microprocessors and used internal RISC core under the old CISC instructions. . . .
Which brings us back to the topic: CISC code being turned into RISC code.
If you read through the above linked conversation, they discuss the fact that Intel does not allow developers direct access to the RISC processors. Intel does not provide any access but x86 CISC code. Their translator remains a black box. Their RISC chips accept code only from that black box. Neither Apple or anyone else can license that black box. It is owned by Intel and will (unless Intel inexplicably change their minds) never let any other company use it.
Result: The CISC to RISC reprogramming problem remains exactly the same as we’ve been discussing. More than merely 5% of CISC code is going to have to be written from scratch to work with ARM RISC processors. Intel of course has a great stake in offering NOTHING to assist that process.
Example: The conversation linked above mentions AVX (Advanced Vector Extensions) and SSE4. Both of these technologies are owned by Intel. Developers are provided with the instruction sets for them, but Intel owns the architecture. You can’t pick up the CISC code that addresses the architecture and apply it to RISC chips. These architectures don’t exist for RISC. Therefore, there are no compilers to translate the CISC code into RISC.
What’s a developer to do to recreate their application for RISC without being able to have access to the architecture their code uses? It would fall to Apple to solve that problem. Apple would have to create or obtain a method of translating the Intel CISC code, which Intel won’t license to anyone else, into ARM RISC code.
I understand the logic behind the wwdc argument for such a big change ( x86 to ARM ). But I think there is not that many important company to be on board .
Apple could have been working with theses guys directly for a while :
Autodesk
Adobe
Microsoft
The foundry
Etc …
althought i don’t know anything about software development Apple seem to be in many many ways in so much better shape to do an architecture transition that they were when they did the ppc to intel one . The have much more control now over it , Don’t they ?
They have not. They have been working with FCPX team, to bring it back into the standards the rest of the industry rely on. Hence why everyone of importance has moved on to Premiere or Avid. Nuke runs much better on a linux machine with the latest GPU then it does on the fastest mac pro with those 4 year old POS GPU’s… Apples hardware dominance is over. Sorry. Metal API is a joke, with only one game supporting it. Apple doesn’t care about VFX. They care about your buying a new phone and a new watch every year.
Well Hello Boys if that is the case because they are sure to be sexed up if true.
Mac excitement returns! Apple fans speculating on new Mac configurations! MDN talking about Macs and future possibilities!
This is starting to sound like the MDN forum of old…
I think Apple is going to announce products that will not be immediately available. I think it will introduce new MacBook and MacBook Pro but I also hope the announce something along the line of an iMacPro or MacPro. I am currently in the market for said machine. The MacPro in it’s current form does not represent good value and Apple knows it.
“The MacPro in it’s current form does not represent good value and Apple knows it.”
Boy is THAT an understatement. In my opinion there should never be a moment when that situation should exist. Even though I know that’s easy for me to say.
I think there will be some updates to their desktop lines. They almost have to at this point. I doubt they will spend much time on it. They will probably show off updated FCPX too. This will most likely be at the beginning of the show and they will quickly move on to the MacBooks. That will be the highlight of the show, and the only thing the press will talk about. Watch the show, re watch it, then visit Apple’s web page to get any information on desktops. I am sure Tim will squeeze in IBM’s new TCO study at the top. Probably at the expense of desktop time on stage.
I want a new Mac Pro as well (an UPGRADABLE machine no less)
looking at the issue at another angle: the MP is built in USA, can Apple afford to abandon building Macs in USA? the political pressure to continue is great as Tim Cook keeps having to defend his position of jobs in USA (latest hits is during the presidential campaigns). In a TV interview a few months ago when Cook was asked about USA manufacturing he NEVER mentioned the MP, probably because he knew it was so out of date, which probably galled him. So here’s just hoping for political reasons (all those ‘ban’ made in China protests etc) T.C will keep MP USA alive (but NOT the Cylinder please, some UPGRADABLE TOWER would be better…)
I grant that the cylinder was not an ideal shape. But had they chosen to they could have made it upgradable using a stacking configuration. The cylindrical cards and housings would have been more expensive but it could have been done and Mac Pro users would have gladly paid the difference. I love Apple but sometimes I just can’t fathom what they’re thinking.
Bring back the CHEESEGRATER!!!
“Benchmarks as fine as….” 3 or 4 year old CPU tech..
what a joke.
The A10 variant for Macs (A10M) could include 8 or more cores at higher clock speeds (compromise on power efficiency and heat production) to increase performance. And Apple’s multi-core GPU embedded in the A10X already kicks 4K/SUHD butt as-is. But Apple could add some more GPU cores, too.
Who knows where this might go. Your skepticism is premature.
Perhaps. Architecture changes need a WWDC.. Because other then all the little app store apps all the big boys would need to re-code for yet another apple processor flavour of the year. Or.. just say F it and keep releasing for workstation class powerhouses running linux and windows.
agree, it wouldn’t make sense to announce an architecture change without giving the devs lead time to make changes. they at least give the summer to update from iOS9 to 10 and from 10.11 to 10.12
Maybe add the A10M along with Intel?
No
That was my thoughts as well….I hope not. Or if they do have a separate line of them along side the intel versions. Moving away from intel and loosing Windows compatibility I think is a mistake if true.
Traditionally, the “Hello” thing has accompanied significant changes of direction rather than incremental upgrades. It seems quite plausible that Apple could be about to announce A-series CPUs in Macs.
I’m convinced that the A series Macs will be a range that sits alongside Intel Macs. Those who need Intel Macs will pay the extra for the Intel CPU, while those who use only applications written for Macs will be able to enjoy cheaper and more power-efficient Macs using A-series CPUs.
I’ve never needed Windows compatibility and would buy an ARM powered Mac as soon as the applications that I use are available for ARM.
The problem with dual architecture options is that Apple has been overrun by beancounters and can’t seem to evaluate each product on its own merits, only by percentage of overall revenue.
Look at the Mac Pro. They generate the least total revenue in a quarterly report even though per-machine margins are high, because it’s a niche market. And even though that niche is important, Apple has ignored it for 3 years (and arguably longer, if you consider that the Mac Pro form factor is unsuited for proper pro work in many ways).
Even Macs in general have suffered, with no major updates to most families in over a year. If we have A-processors in the least expensive offerings, which traditionally are what sells the most, the beancounters can say “hey look, the Intel line isn’t selling well” and convince the execs (if the execs aren’t the very shortsighted beancounters in the first place) that not enough customers are interested in Intel compatibility, leading to either higher-end Macs languishing, or cut altogether.
I’m convinced it is going to be a big announement. THey are obviously channeling other historic announcements. Add that to the time the Mac has languished plus the effort Apple is known to put into their products. I’m expecting a lot.
That was my thoughts as well….I hope not. Or if they do then have a separate line of them along side the intel versions. Moving away from intel and losing Windows compatibility I think is a mistake if true. Also the amount of time between upgrades is another sign of major changes coming.
iPhone and iPad with A10X. Able to plug into the new Thunderbolt Display. 256Gb storage option on these devices allows them to become the primary computing device.
Walk around with your primary and only computer in your pocket. Full size apps when you plug into the screen at home but work as normal iOS variants when screen unplugged.
Switching processors brings up the issue of getting old code to run on new processors. How does Apple pull off a processor switch once again? Is there a version of Rosetta that translates x86 to ARM instructions?
Apple does not need to use the Rosetta approach again. Apple probably already has Xcode ready to compile for ARM. Same source code – multiple executable targets.
No actually. You still don’t understand the difference writing for CISC CPUs and RISC CPUs. It is far from as simple as you pretend. I keep pointing the way and if you don’t go there you end up spreading nonsense about this situation.
As a software engineer and previously a RISC processor ALU designer I disagree.
Yes, there will always be complications through any transition, but I disagree that this is an impediment or would stop Apple from doing whatever is most compatible with where they see computing in 5-10 years.
A focus on the product results in growth for Apple down the road, even as it often generates pain for users in the short run.
On the technical side, most code can just be recompiled. Not all, but most as in 99.5%. Of course the 0.5% that must be rewritten is often very foundational/complex low-level processor/device specific code. But it is in the minority.
I would see it as Apple’s job to provide as much low-level transition support as possible (BLAS libraries, make real gurus available for developers, etc.) while making most code just recompile seamlessly.
Apple has the advantage of owning their whole tool, software, OS and hardware stack, and even an optional software store for handling sales, updates, etc. They won’t move until they think everything is lined up right.
And I agree with the Intel hold outs, Apple will continue shipping Intel Macs. It will be the low end that gets A-series chips first where performance and Windows compatibility are not issues for most people.
I profoundly disagree about the required code rewrite. Don’t foist nonsense please. People tend to believe it these days. The usual advice: Study CISC vs RISC. You clearly have not.
If you actually read what he wrote, he agreed that most of the code could simply be recompiled. Rather, a small percentage of code would have to be rewritten. As someone who has ported applications across architectures (same OS), I agree with his assessment and would suggest that arguing this point simply demonstrates that you don’t know what you’re talking about. Generically speaking there could be endian issues to deal with, but even that isn’t likely. Your CISC vs RISC comments are pretty comical. If you’re writing in Swift, how exactly do you expect to be impacted by this? If you’re writing in assembler, it would be an issue, but then you’d need a complete re-write anyway.
Go read the reply I wrote to ‘Michael’. I am not going to laugh at you for your being wrong, so don’t worry.
To be clear, I’m not agreeing with KingMel’s position that something like Rosetta wouldn’t be needed. It would. Rather, I’m laughing at your explanation as to why it would be needed. Perhaps you might want to rethink your reasoning before commenting further. Hint: CISC vs RISC really has nothing to do with it.
Thanks for making factual arguments. I agree with them.
Yes, porting low-level code will take work. Apple will no doubt take that into account and work to make that as easy as possible. Their may be some emulation, but certainly tools to help with that.
But don’t expect Apple to hold off from what it sees as a long term advantage (if it does see owning its Mac processors as advantageous) due to the short term pain for a minority of developers.
Product comes first for Apple, not developers. They have truly screwed developers many times. Developer pain is not hight on the priority list for for Apple as they have demonstrated many times.
Yes, I agree with your position that Apple will do what Apple thinks is best, regardless of developers needs, etc. I don’t believe there is much of a technical barrier for Apple at this point. There are other things to consider though. Chip development is expensive. Apple certainly “can” build a better chip than Intel based on ARM’s instruction set. I don’t think that’s in question. The question is about benefits.
It makes sense for Apple to do custom silicon for their iPhones because the iPhone is Apple’s “bread and butter” and because Apple ships products in ridiculously high volumes with these chips. Especially when you include iPads, iPod Touch, AppleTV, etc. The volume of Mac shipments is very low by comparison and would hardly justify the cost of a custom CPU, etc. Further, the ability to run Windows apps (boot camp or virtual PC) has proven to be an overall benefit for Apple in selling Macs. It’s not clear that creating a custom Mac ARM based CPU would be worth it.
Derek, I would rarely disagree and would almost never be read to you.
But f#ck you. I have designed chips. I know the difference between RISC and CISC far better than you probably do as have designed instruction sets for both.
Your faceless ad hominem attack is not like you, so I look forward to more sensible discussions in the future.
1) RISC to CISC instruction set conversion is easy. Their are profound differences in how they pipeline and branch predict, etc., for performance but the actual translation for equivalent functional behavior is EASY when conversion is from C or another higher language in both cases.
2) x86 is not just CISC but CISC laden with enormous design debt. It is actually very inefficient in terms of silicon area compared to what the same CISC functionality would have if designed in modern times.
3) RISC by definition has very little design baggage as there is so much less to complicate.
Get a f@cking education on RISC and CISC – by which I mean actually go through the process of designing instruction sets and fabbing processors. I HAVE.
And when you communicate a view BACK IT UP with objective points. I continually see you making CISC > RISC comments without the least bit of factual information.
You are wrong to start with. But when you don’t back up your points objectively you are not even wrong.
Derek, I apologize for coming on strong, but not for strongly disagreeing with you from real experience.
Normally I agree with just about every view you express.
As I posted elsewhere today, I’m going to sit back and snooze through this nonsensical subject into the future. Have fun watching nothing happen, for the reasons I’ve been jabbering for a year now, and more… 💤💤
I am not looking forward to transitioning to a new cpu architecture again. And, I don’t think that is what is going on. I think it more likely that we will see a new device. One that is between the ipad and a macbook. I believe MDN has said that these two platforms will continue to merge/blend over time.
This had better be good. I hope Apple focused on the four: MacBook Pro, Mac mini, iMac, and Mac Pro. We need two, MacBook Pro, and Mac mini.
Wouldn’t such an announcement be more of a WWDC thing?
Obviously yes. Thank you for pointing that out. Moving the Mac to A-Series RISC CPUs would cause outrage and uproar among ALL Mac developers. That’s one of the several reasons this is a Nonsense Meme.
Nobody complained about developing for ARM/iOS/iPhone/iPad, or for watchOS. As long a there is a compelling advantage for customers to have A-series laptops (say cheaper, faster for the money, longer batter life, etc) developers will be fine. I am one of them.
What would be bad would be if Apple announced A-series laptops without a big customer-centric reason for getting them. That would be a disaster for Apple and developers would just ignore the new machines.
Apple knows this. They are not just going to swap CPUs. If they swap CPUs it will be to deliver something obviously better for a well defined and large group of customers.
Oh, look. I get to repeat myself. Sorry, too bored to bother.
I seriously doubt there will be a switch to ARM for Macs. There is a lot of reasons not to it, with few reasons to do it. This has been in the press for years, like the Apple HDTV that everyone knew was going to come “next” year.
Like most Apple events the press will report on some wild ideas they think Apple will do, then complain that Apple did not do what they made up. They will spend little space on what the new Apple product can do because that would mean real journalism. The ARM Mac is another shot in the dark so if it does happen then people will say I right, see how smart I am.
Wish Apple would just buy an x86 license and take their CPU knowledge and make their own chips for the mac. If they did that, they wouldn’t need to shift every mac customer/developer to a new architecture, and they could incorporate all they know about their A series CPUs in the new x86 chip.
If Jobs were still running Apple, I think we would have already seen the switch away from Intel. There is really no major reason to stay with Intel anymore, when they have their own chip design. The performance is clearly there, the ability to ramp up manufacturing as well, their chip design team is extremely successful at making major strides each year.
Let us not forget; the move to Intel wasn’t so that the Mac could run Windows (although at the time, with significantly lower market share, as well as strong marketing push, it was a bonus). It was because IBM/Motorola weren’t delivering. Today, Intel’s roadmap isn’t any more ambitious than Apple’s own for CPUs. Bringing everything under their own control reinforces the philosophy of designing your own software and hardware.
There are very few Mac users who still run Windows. Their numbers are negligible; probably fewer than those who still use optical disc on a Mac. Apple has already eliminated that optical disc (as they did many other features and ports).
Again, if Jobs were still alive, he would have already moved over, and pushed everyone (Adobe, Microsoft, other big players) towards the new architecture. However, without Jobs, I’m not so sure.
I’m not so sure. Were he alive today, Steve Jobs would have changed along with shifting market preferences. His predilection for guillotining established tech standards in favour of proprietary new-age gimcracks would have by now matured into a real appreciation for the angry pain of transition suffered by so vast a number of sorrowful pro users. His humanity had increased to the point of actual compassion by the time he passed. Rather than drag everyone screaming into the future, as he once tried to do, he’d have learnt not to blow off a hat-in-hand emailer with an abrupt “you’re holding it wrong,” but to turn around and yell at his engineers that “you’re designing it wrong.”
Please cite your source for the statistics about the vanishing base of Mac users who run Windows. It seems too careless an assumption, and not consistent with my experience. You once pointedly cited studies of gunshot death rates by country to argue for the efficacy of gun control. Is not the pattern of OS usage as worthy of statistical support? Granted, nobody running Windows actually dies. They only wish they had… 😬
A dual or quad A10x powered Mac would be awsome!
The days of intel consumer computers is nigh, so it may not be this year but far sooner than people think. Old technology that is kept around for the some purpose of lack of vision and it being today’s bread and butter product pushed by the sales staff, is ALWAYS the death knell of EVERY company that has ever existed.
Intel doesn’t seem to be pivoting the same way that IBM or Microsoft has. It looks like it will be going the way of Smith Corona.
The only way Apple would move to ARM based processors for the Mac line would be either a) to significantly reduce the hardware build cost of the Mac, so they could enter a lower price tier without giving up margins, or b) they want to migrate everything to IOS, and that would be a temporary means to start moving in that direction. They could choose either reason, but I do not believe they will.
I suggested years ago that Apple build an IOS compatibility module into MacOS so that IOS Apps could run on both platforms, but nobody listened back then.
Sources tell me also iMac may not even get refreshed in this announcement.
Dear, Lord, please bring back the old Apple….
I wonder whether the ‘space’ like design of the invite shows an upgrading of cloud infrastructure, maybe some kind of A powered Macs to compete with Chromebooks, especially for areas like education where Apple marketshare has dropped to about 10% and Chromebooks climbed to 50% (the education market is a very important area for Apple as besides revenue it trains new users).
An iCloudBook would be a good idea for education, if Apple were able to hit near Chromebook pricing. There is also a segment of consumers whose computer activities are already Internet-based, meaning they use a web browser, email and social media. So requiring an active internet connection wouldn’t be a huge change from their use pattern. That would be interesting for the low-end, if Apple decided to play in that space. Typically Apple has fancied itself a Mercedes and not a Kia.
‘Hello Again’ means Apple ‘A’ series-powered Macs
No it doesn’t. Not gonna happen. I’ve been over the details of why this is a nonsense meme at least 10 times and I won’t be going over it again. Go look up ‘CISC’ and ‘RISC’ and learn why for yourself. Just don’t bother me with your lack of comprehension. Thank you.
You do know that Intel’s CPUs are in fact RISC cores and have been for many, many years? “CISC” instructions are broken down (transcoded) before being passed along to the actual RISC core to be processed.
… And here’s a very good conversation about that situation:
x86 – Why does Intel hide internal RISC core in their processors?
Starting with Pentium Pro (P6 microarchitecture), Intel redesigned it’s microprocessors and used internal RISC core under the old CISC instructions. . . .
Which brings us back to the topic: CISC code being turned into RISC code.
If you read through the above linked conversation, they discuss the fact that Intel does not allow developers direct access to the RISC processors. Intel does not provide any access but x86 CISC code. Their translator remains a black box. Their RISC chips accept code only from that black box. Neither Apple or anyone else can license that black box. It is owned by Intel and will (unless Intel inexplicably change their minds) never let any other company use it.
Result: The CISC to RISC reprogramming problem remains exactly the same as we’ve been discussing. More than merely 5% of CISC code is going to have to be written from scratch to work with ARM RISC processors. Intel of course has a great stake in offering NOTHING to assist that process.
Example: The conversation linked above mentions AVX (Advanced Vector Extensions) and SSE4. Both of these technologies are owned by Intel. Developers are provided with the instruction sets for them, but Intel owns the architecture. You can’t pick up the CISC code that addresses the architecture and apply it to RISC chips. These architectures don’t exist for RISC. Therefore, there are no compilers to translate the CISC code into RISC.
What’s a developer to do to recreate their application for RISC without being able to have access to the architecture their code uses? It would fall to Apple to solve that problem. Apple would have to create or obtain a method of translating the Intel CISC code, which Intel won’t license to anyone else, into ARM RISC code.
Advanced Vector Extensions
(Aren’t I nice to do people’s homework for them? I hate doing other people’s homework!)
I understand the logic behind the wwdc argument for such a big change ( x86 to ARM ). But I think there is not that many important company to be on board .
Apple could have been working with theses guys directly for a while :
Autodesk
Adobe
Microsoft
The foundry
Etc …
althought i don’t know anything about software development Apple seem to be in many many ways in so much better shape to do an architecture transition that they were when they did the ppc to intel one . The have much more control now over it , Don’t they ?
They have not. They have been working with FCPX team, to bring it back into the standards the rest of the industry rely on. Hence why everyone of importance has moved on to Premiere or Avid. Nuke runs much better on a linux machine with the latest GPU then it does on the fastest mac pro with those 4 year old POS GPU’s… Apples hardware dominance is over. Sorry. Metal API is a joke, with only one game supporting it. Apple doesn’t care about VFX. They care about your buying a new phone and a new watch every year.
OMG who would run Maya on a Mac? that would be so painful.. so much waiting around..
And based on the announcement, I’m assuming The Cars will be the musical entertainment for the event?
🙂