Bloomberg: Apple plans 15-Inch MacBook Air, new 12-Inch laptop.

[Colstan]

So, Gurman is reading chicken entrails again. According to his long-winded article, Apple is planning a 15-inch MacBook Air and a new 12-inch laptop for 2023 or early 2024. He also reiterates some things he already published:

The M2 Max chip in the next high-end MacBook Pros will include 12 main processing cores and up to 38 graphics cores, up from a maximum of 10 processing cores and 32 graphics cores in the current models. New versions of the Mac mini and a revamped Mac Pro are also in testing within Apple, Bloomberg has reported. And the company is already working on an M3 chip destined for a future iMac and other products.

So, the M3 is in development. I'm glad he's here to tell us these things.

One item that the article highlights is that the Mac business has really taken off since switching to Apple Silicon. All the doomsayers that claimed losing x86 and Boot Camp would kill the Mac were quite wrong.

 

[Cmaier]

So, Gurman is reading chicken entrails again. According to his long-winded article, Apple is planning a 15-inch MacBook Air and a new 12-inch laptop for 2023 or early 2024. He also reiterates some things he already published:



So, the M3 is in development. I'm glad he's here to tell us these things.

One item that the article highlights is that the Mac business has really taken off since switching to Apple Silicon. All the doomsayers that claimed losing x86 and Boot Camp would kill the Mac were quite wrong.

Yeah, the boot camp ninnies were my favorite at the other place.

So, for the mac pro, we are looking at, what, 48 cores? Geekbench 5 of 2000 single thread and 63,000 multi thread?

 

[Colstan]

Yeah, the boot camp ninnies were my favorite at the other place.

As you've stated previously, it's highly unlikely that the Mac Pro will use external DIMMs for system memory. I expect the next great downfall of the Mac will be the fact that the Apple Silicon Mac Pro won't be able to support up to 1.5TB of RAM. The gnashing of teeth and substantial angst will, of course, come from people who can barely utilize 16GB, if that.

So, for the mac pro, we are looking at, what, 48 cores? Geekbench 5 of 2000 single thread and 63,000 multi thread?

If you believe Gurman, from a previous article back in March:

• M2: eight CPU cores and nine or 10 graphics cores
• M2 Pro: 12 CPU cores and 16 graphics cores
• M2 Max: 12 CPU cores and 32 graphics cores
• M2 Ultra: 24 CPU cores and 48 or 64 graphics cores
• M2 Extreme: 48 CPU cores and 96 or 128 graphics cores

Your performance estimates are as good as mine. We can extrapolate more once benchmarks on the M2 are released. If the M1 Ultra is any indication, the memory bandwidth on the M2 "Extreme" is going to be obscene.

 

[theorist9]

[Gurman quote:]

"The M2 Max chip in the next high-end MacBook Pros will include 12 main processing cores and up to 38 graphics cores, up from a maximum of 10 processing cores and 32 graphics cores in the current models."

38 GPU cores for the Max seems a curious number

Given that Apple says the base M2 has 10 GPU cores, and the number doubles with each step for the M1, I would have expected something like this for the M2:
M2 base 10
M2 Pro 20
M2 Max 40
M2 Ultra 80

 

[Cmaier]

As you've stated previously, it's highly unlikely that the Mac Pro will use external DIMMs for system memory. I expect the next great downfall of the Mac will be the fact that the Apple Silicon Mac Pro won't be able to support up to 1.5TB of RAM. The gnashing of teeth and substantial angst will, of course, come from people who can barely utilize 16GB, if that.


If you believe Gurman, from a previous article back in March:

• M2: eight CPU cores and nine or 10 graphics cores
• M2 Pro: 12 CPU cores and 16 graphics cores
• M2 Max: 12 CPU cores and 32 graphics cores
• M2 Ultra: 24 CPU cores and 48 or 64 graphics cores
• M2 Extreme: 48 CPU cores and 96 or 128 graphics cores

Your performance estimates are as good as mine. We can extrapolate more once benchmarks on the M2 are released. If the M1 Ultra is any indication, the memory bandwidth on the M2 "Extreme" is going to be obscene.

RAM will be interesting. I figure there are a couple of possibilities. Either you assume 24GB per M2 is a structural maximum, and you assume it works like M1, in which case M2 Max is up to 96GB, M2 Ultra is 192GB, and M2 Extreme is 384GB. Or you assume that the only thing keeping M2 at 24GB instead of 32GB, given that once they went above 16GB they had to add an extra address bit already (which means they can address up to 32GB), is physical limitations in RAM density (or cost limitations - maybe higher density RAM is more expensive or the packaging for the chip would be too expensive), in which case Extreme may go “cost is no object” and be sold with up to 512GB.

It will be easier to tell once we see M2 Max.

I think M2 Extreme ends up being faster than Core i9-12900K in single core, and around the speed of AMD‘s current 128-core EPYC’s on multi-core. So very fast, at a tiny fraction of the power consumption of those. Still likely won’t be the fastest chip on the market by the time it comes out, but it may be close enough.

 

[theorist9]

So, for the mac pro, we are looking at, what, 48 cores? Geekbench 5 of 2000 single thread and 63,000 multi thread?

In doubling the number of CPU cores from the M1 Max Studio to the M1 Ultra Studio, the GB5 MT score increases by 1.9x. So if we assume another 1.9x increase in going from the Ultra to the Extreme, and factor in Apple's stated 18% increase in MT CPU performance between the M1 and M2 base chips, we get:
23366 x 1.9 x 1.18 = 52,000

However, much of that 18% increase in MT performance might be due to significantly increased performance of the efficiency cores in the M2 vs the M1, in which case the increase in SC performance could be <18%. In that case, since the Ultra has a higher ratio of Big:little cores than the base chips, the GB MT score may be <50,000 (and the M2's SC score could be < 2000).

 

[Cmaier]

In doubling the number of CPU cores from the M1 Max Studio to the M1 Ultra Studio, the GB5 MT score increases by 1.9x. So if we assume another 1.9x increase in going from the Ultra to the MP, and factor in Apple's stated 18% increase in MT CPU performance between the M1 and M2 based chips, we get:
23366 x 1.9 x 1.18 = 52,400

However, much of that 18% increase in MT performance might be due to significantly increased performance of the efficiency cores in the M2 vs the M1, in which case the increase in SC performance increase could be <18%. In that case, since the Ultra has a higher ratio of Big:little cores than the base chips, the GB MT score may be <50,000 (and the M2's SC score could be < 2000).

Could very well be. I estimated 2000 for ST and multiplied by the number of cores then de-rated by a factor of .66 for scaling inefficiency, which is around what M1 Ultra’s factor might be.

 

[Herdfan]

Yeah, the boot camp ninnies were my favorite at the other place.

I understand purists hating boot camp, but it brought people like me who still needed the ability to run a windows program into the MAC fold. Now that the program I was using is now online, I no longer need it. But without it, I would have never bought a Mac in the first place. It helped expand the brand.

 

[Colstan]

I understand purists hating boot camp, but it brought people like me who still needed the ability to run a windows program into the MAC fold. Now that the program I was using is now online, I no longer need it. But without it, I would have never bought a Mac in the first place. It helped expand the brand.

It's not about being purists, it's about Boot Camp no longer being needed. I shared these numbers a half-dozen times over at MR, but haven't done it here, so I might as well. Apple Insider has contacts within the Mac repair and service industry. According to them, in 2010, 15% of Macs had Windows installed using Boot Camp. By mid-2020, before the transition to Arm, only 2% were using Boot Camp. If that trend continued, by the time Apple Silicon became available across the line (sans Mac Pro), Boot Camp wasn't even a rounding error. Comparatively, an informal poll showed that 35% of Apple Insider's readership was using Boot Camp in 2020.

We need to remember that tech nerds like us don't represent the average user. We're an important minority, particularly in promoting Apple's ecosystem through indirect means, but don't reflect the overall market. Also, I have no issue with Boot Camp. I still use it on my Intel Mac mini to play an occasional Windows game. I think it would be great to see Boot Camp return for Apple Silicon Macs, but I'm realistic enough to know that's not going to happen. Microsoft has liscensing issues with Qualcomm. Apple has no interest in spending engineering resources writing custom Windows drivers, neither does Microsoft, and it's questionable if Apple would provide them with documentation. Apple and Microsoft continue to be competitors; something that we often forget. Boot Camp was a product of its time, an easy "freebie", and helped entice reticent PC users to the platform. The Mac has grown past the need for such crutches and can stand on its own.

 

[Cmaier]

I understand purists hating boot camp, but it brought people like me who still needed the ability to run a windows program into the MAC fold. Now that the program I was using is now online, I no longer need it. But without it, I would have never bought a Mac in the first place. It helped expand the brand.

I understand that some tiny percentage of people need it for whatever reason. But MR is full of people who claimed that eliminating Bootcamp would be the end of the Mac, because so many people supposedly require it.

Of course that‘s false.

And as I pointed out repeatedly, you *gain* access to much more software than you lose - between iPad/iOS apps that already run on Mac and the fact that unifying the architecture will mean that much more software will be written for Mac because it will also run on the much larger iOS/iPadOS ecosystem, the software library for Mac has already grown substantially compared to what was lost.

Not to mention, that if Windows eventually allows licensing of Windows Arm on non-Microsoft devices (which they‘ve taken steps toward already), at some point Windows will be back, even if it‘s only in a VM.

 

[theorist9]

It's not about being purists, it's about Boot Camp no longer being needed. I shared these numbers a half-dozen times over at MR, but haven't done it here, so I might as well. Apple Insider has contacts within the Mac repair and service industry. According to them, in 2010, 15% of Macs had Windows installed using Boot Camp. By mid-2020, before the transition to Arm, only 2% were using Boot Camp. If that trend continued, by the time Apple Silicon became available across the line (sans Mac Pro), Boot Camp wasn't even a rounding error. Comparatively, an informal poll showed that 35% of Apple Insider's readership was using Boot Camp in 2020.

We need to remember that tech nerds like us don't represent the average user. We're an important minority, particularly in promoting Apple's ecosystem through indirect means, but don't reflect the overall market. Also, I have no issue with Boot Camp. I still use it on my Intel Mac mini to play an occasional Windows game. I think it would be great to see Boot Camp return for Apple Silicon Macs, but I'm realistic enough to know that's not going to happen. Microsoft has liscensing issues with Qualcomm. Apple has no interest in spending engineering resources writing custom Windows drivers, neither does Microsoft, and it's questionable if Apple would provide them with documentation. Apple and Microsoft continue to be competitors; something that we often forget. Boot Camp was a product of its time, an easy "freebie", and helped entice reticent PC users to the platform. The Mac has grown past the need for such crutches and can stand on its own.

I agree about the general trend you're reporting (a decreasing need to run Windows), and also the general sentiment (that, for most users, BootCamp is not needed).

But, quantitatively, I don't think one should calculate Windows-on-Mac use via Boot Camp figures alone, since it's also done with Parallels, VMware Fusion, and Virtual Box, as well as other minor players. Just prior to the AS transition, Parallels claimed 7 million users. If Parallels' number is legit (I don't know if it is), and we assume an installed base of 100M Intel Macs, and add a few percent for all other Windows-on-Mac vehicles, that gets us to about 10%.

Plus, while the Mac can increasingly stand alone generally, it should be recognized that there are still large islands where MacOS simply can't be used, mostly within large businesses and institutions that have legacy PC-based software that would cost millions to change. [E.g., I suspect the entire financial/fund administration/HR portion of many large universities is PC only.] I think this is the largest remaining market Macs haven't yet penetrated. Then again, you're not going to penetrate that market at a large scale with virtualization. Large-scale penetration won't happen until that legacy software (or whatever replaces it) is written to run natively on the Mac.* So the main value of virtualization will be for the (fewer) number of workers who have a personal choice of what computer to use.

*Actually, even then you might not penetrate it. SOP in those markets is a cheap Dell PC box attached to two cheap ~100 dpi 24" monitors (~$200 each). You could replace the Dell with a Mini but, compared with Windows, current versions of MacOS look really bad (headache-inducing, IMO) on 100 dpi monitors. MacOS really needs Retina in order to look good, and that makes for two really expensive external monitors. I suppose a 24" iMac plus a cheap 24" 4k external would be close enough, but they might balk at the up-front cost.

 

[Colstan]

But, quantitatively, I don't think one should calculate Windows-on-Mac use via Boot Camp figures alone, since it's also done with Parallels, VMware Fusion, and Virtual Box, as well as other minor players.

Sure, there's a reason that Apple has showcased Parallels running Linux VMs, which is really just a wink and a nod toward those who will be running Arm Windows through virtualization, even if Microsoft doesn't currently allow it through licensing. They also worked with CodeWeavers to allow 32-bit x86 apps to run through CrossOver using Rosetta 2. Apple knows that the demand is there, but they aren't going to divert valuable engineering resources and talent toward a niche among a niche by implementing Boot Camp. They will assist others in alternative implementations.

Here is a direct quote from Craig and his fabulous hair.

Most of the folks who were grousing over the loss of Boot Camp over at MR were either people who had some oddball app that didn't work well through VMs, or gamers who aren't happy about the selection of computer games on the Mac. The former are in an unfortunate situation, but don't matter in overall market share numbers, and the latter are being addressed with Metal 3, which Apple hopes will improve game selection. (But, you know, nothing makes the MR crowd happy.)

 

[Colstan]

Mark Gurman, a formerly reliable leaker, and heir apparent to Digitimes, is now being contradicted. One of the last remaining quality analysts, Ross Young, is skeptical about a 12-inch MacBook being in the works:

Apple's strategy for notebooks is currently 13" and larger. Companies in the MacBook Pro display supply chain we talked to are not aware of it.

While it is certainly possible that a 12-inch MacBook could be on Apple's roadmap, Young has seen no evidence of it within the supply chain. Ross Young has been hyper-focused on display panels, and is perhaps the most reliable, accurate leaker for future iMacs, MacBooks, and Apple monitors. He was incorrect about the last 27-inch panel being used for an iMac, because it had a fully integrated A13 and 64GB of storage, which is what we now know as the Apple Studio Display. While he was incorrect about the panel's intended purpose, he was correct about its approximate release date and general specs.

Whether Gurman is proven accurate in the long-term, the quality of his rumors has suffered, perhaps due to Apple cracking down on those doing the leaking. Kuo also missed boarding the kayak with his prediction that the new MacBook Air would use an M1. They can always use the "plans change" excuse in the future, crowing about the things they get right, while ignoring those that they get wrong, bloviating about their skilled augury. I think Ross Young is now the most reliable leaker, because his contacts are specifically within the display panel manufacturers. Hence, his information isn't as broad, but is far more reliable.

Per the [Bloomberg] report, the company is also working on a 15-inch MacBook Air based on the recent redesign. Both the 12-inch MacBook model and the 15-inch MacBook Air would sport some type of M2 chipset, and could be ready for a launch by late 2023 or early 2024.

So, if anyone is looking for a 12-inch Apple Silicon MacBook, it's not likely to happen anytime soon, if at all, assuming Ross Young is to believed.

 

[theorist9]

Ross Young, CEO of Display Supply Chain Consultants, whom you mentioned above, and whose predictions are considered reliable because they are based on display components that have gone into production, predicts a 15" Air will be released in April 2023:

A new 15-inch M2 MacBook Air could be here as soon as April

Ross Young of Display Supply Chain Consultants says the display has entered production.

www.macworld.com