The new Mac mini is almost certainly coming

[Spectrum]

In general, I see wider variation than you: 86-96C when encoding, with most cores in the 90s, and the 'yes' test is 87-97C, with 3 cores in the 90s, and 3 in the 80s.

Thanks! Then I think the only explanation is that your heatsink/fan is indeed working more efficiently - allowing the CPU to stay cooler despite using more power.

In any case, since I don't do encoding all day, the performance is perfectly acceptable for my workloads. The i7 does still get a CPU boost on single/double treaded tasks (up to 4.5-4.6 Ghz) relative to the i5 (up to 4.0-4.1). It is only in a multi-core task that the gains relative to your i5 seem marginal. 👍

 

[Krevnik]

Thanks! Then I think the only explanation is that your heatsink/fan is indeed working more efficiently - allowing the CPU to stay cooler despite using more power.

In any case, since I don't do encoding all day, the performance is perfectly acceptable for my workloads. The i7 does still get a CPU boost on single/double treaded tasks (up to 4.5-4.6 Ghz) relative to the i5 (up to 4.0-4.1). It is only in a multi-core task that the gains relative to your i5 seem marginal. 👍

I've almost never seen 4.1Ghz on the i5.

I did get a chance to play with an i7 and put it through my test loads, and got results a lot closer to yours. 10% faster while using less power when in Handbrake (~60-70W). For compiles, also about 10% faster, higher CPU frequencies, and higher CPU loads. Also managed to see some spikes well above 100W. Higher idle power draw though by a watt or so. Still quite good.

I'm thinking that comparing the i5/i7 in terms of clock speed and power draw at 100C just doesn't work. HT throws too much of a wrench into the works.

 

[dogslobber]

In just three weeks time we'll be at the point where we've not had a new Mini released this decade. Who knew?

 

[Duncan68]

In just three weeks time we'll be at the point where we've not had a new Mini released this decade. Who knew?

Have there been any estimates of the 2018 Mac mini sales numbers? If they’re low enough, maybe this is the last one?

 

[dogslobber]

Have there been any estimates of the 2018 Mac mini sales numbers? If they’re low enough, maybe this is the last one?

Are you in the correct thread?

 

[sublunar]

I would reckon the next CPU update of interest to Apple for the Mini would be Comet Lake S - where everything gets multi threading and 8/10 core CPUs (with 16/20 threads) become mainstream 65w parts in the second half of 2020.

It's the final 10Nm refresh until Ice Lake (7Nm) arrives the following year.

Coffee Lake Refresh is available now but is only a mild speed bump of existing CPUs and would likely not make much sense if Apple can wait till October 2020 for the iMac refresh as well.

 

[EightyTwenty]

I would reckon the next CPU update of interest to Apple for the Mini would be Comet Lake S - where everything gets multi threading and 8/10 core CPUs (with 16/20 threads) become mainstream 65w parts in the second half of 2020.

It's the final 10Nm refresh until Ice Lake (7Nm) arrives the following year.

Coffee Lake Refresh is available now but is only a mild speed bump of existing CPUs and would likely not make much sense if Apple can wait till October 2020 for the iMac refresh as well.

I believe Ice Lake is 10nm, which are only available in laptops today. I don’t think Intel is coming out with 7nm chips anytime soon.

 

[PickUrPoison]

Have there been any estimates of the 2018 Mac mini sales numbers? If they’re low enough, maybe this is the last one?

You’ll never get any model-specific sales data from Apple. But we’re probably six to nine months away from the earliest refresh that could be expected.

The only Mac that (usually) gets yearly updates, and sometimes even more often, is MBP. All other models are on approx. a two-year refresh cycle, which are usually driven by CPU upgrades. I expect iMac Pro, Air and mini next year; maybe an iMac redesign as well.

 

[PickUrPoison]

AMD has a APU but the next Mini will be ARM based, I think we can all agree with that. I would not buy a Mac at this point in time. Personally, I would hold off for a year, if at all possible.

I’m not sure how many would agree with that, it’s currently a 65W CPU. I’d expect they would transition the Air first (and in my fantasy world resurrect the MacBook form factor, though I’d prefer a 13/14” display). Then move to the MBP, if they’re dumping Intel there too. Then mini and iMac.

Could the entire lineup cut over more quickly? Possibly, but I think it’ll take two to three years, with iMac Pro and Mac Pro remaining Intel machines.

 

[sublunar]

I’m not sure how many would agree with that, it’s currently a 65W CPU. I’d expect they would transition the Air first (and in my fantasy world resurrect the MacBook form factor, though I’d prefer a 13/14” display). Then move to the MBP, if they’re dumping Intel there too. Then mini and iMac.

Could the entire lineup cut over more quickly? Possibly, but I think it’ll take two to three years, with iMac Pro and Mac Pro remaining Intel machines.

Looking around at the prospect of ARM in a Mac, I'd say it's more likely that Apple keeps the platforms separate for now given what they are doing with the iPad Pro which easily traverses into MacBook pricing territory when you get the larger screen sizes and storage.

What's more likely for me is an A13X powered AppleTV 4K 'Pro' that can play Apple Arcade games as a games console - perfect timing as next generation games consoles are coming next year.

 

[PickUrPoison]

Looking around at the prospect of ARM in a Mac, I'd say it's more likely that Apple keeps the platforms separate for now given what they are doing with the iPad Pro which easily traverses into MacBook pricing territory when you get the larger screen sizes and storage.

What's more likely for me is an A13X powered AppleTV 4K 'Pro' that can play Apple Arcade games as a games console - perfect timing as next generation games consoles are coming next year.

An ARM Mac would run MacOS, so the software platforms would still be separate. But running Windows with acceptable performance wouldn’t be possible afaik, since ARM wouldn’t be fast enough to emulate x64 without a significant performance hit. Lots of cores are great for certain workloads, but for most applications single-thread ips is key.

re: an AppleTV with A13X, there have been rumors of a AppleTV refresh, but no rumors of A13X. iPad Pro looks to be coming out next fall, so it’ll likely have A14X. So no A13X, just as there was no A11X.

 

[Krevnik]

An ARM Mac would run MacOS, so the software platforms would still be separate. But running Windows with acceptable performance wouldn’t be possible afaik, since ARM wouldn’t be fast enough to emulate x64 without a significant performance hit.

If you are talking about Bootcamp, there's always Windows on ARM. Depends on if Microsoft will sell you a license to the SKU they use on the Surface Pro X. Emulation performance of x86 apps still doesn't seem to be great though. But that's not too surprising since the X is more tablet than high-end laptop. Apple would take similar hits running x64 software on an ARM64 system a la Rosetta, TBH.

Lots of cores are great for certain workloads, but for most applications single-thread ips is key.

This is true for legacy applications, but even those are getting more multithreaded as time goes on. I'd need a citation of some real data before I accept any claim one way or another these days. Things have changed a lot even in the last 5 years, precisely because single-thread IPS is not where the big gains are anymore.

It also depends quite a bit on the workflow. Office is still very single-threaded, but it's not exactly like people are going to be buying an i9 for office productivity, as it's a waste of money despite the higher single-thread IPS. My workflow is fairly scalable to the number of cores available (to a point), but I admit that not everyone has my workflow, but all the more reason for needing to lean on data for this one.

 

[giggles]

Waiting for a intel 8-core/10-core Mini with “T3” arm chip, 8TB max ssd, U1 ultrawideband for enhanced Continuity, pcie 4.0, thunderbolt 4, Usb 4, hdmi 2.1, DisplayPort 2.0 and native support to the XDR 6K60Hz HDR display and single-tile 8K60Hz HDR displays.

Anything less I don’t see the point over the current one.

 

[sublunar]

An ARM Mac would run MacOS, so the software platforms would still be separate. But running Windows with acceptable performance wouldn’t be possible afaik, since ARM wouldn’t be fast enough to emulate x64 without a significant performance hit. Lots of cores are great for certain workloads, but for most applications single-thread ips is key.

re: an AppleTV with A13X, there have been rumors of a AppleTV refresh, but no rumors of A13X. iPad Pro looks to be coming out next fall, so it’ll likely have A14X. So no A13X, just as there was no A11X.

Running Windows is key here, an ARM Mac could not, as you say, run windows. I wouldn't like to confuse matters with some Macs incapable of running Windows natively or (worse) with poor performance compared to 'proper' Windows.

Instead, rather than trying to compete on a different platform, I could see Apple developing iPad OS further with some new hardware products.

We're not far away from seeing iPad Pros crossing into MacBook territory in terms of performance. I'd like to see if Apple introduce an iPad Pro set up like a laptop rather than a compromised third party solution that relies on people purchasing a keyboard case and using bluetooth connectivity for keyboards.

What if Apple wanted to try a 2-in-1 iPad with a scissor keyboard and USB-C port(s)?

The MacBook is already off sale, so there could seriously be a gap in the $899-$999 area where an ultraportable 'laptop' running iOS apps with a touch screen could be interesting.

iPad Pros have not always come out in October - there have been introductions at WWDC and in March. The last 'Pro' refresh was October 2018 but there have been consumer refreshes in March and September 2018.

I could see demand for cpus for iPhone hottest in Q4 and Q1 - so introducing 'pro' devices in Q2 and Q3 spreads the demand around the year and looks healthier for the balance sheet.

In this case though, such a device would clearly be a candidate for an October release (and perhaps therefore an A14X) not least because of the fact that the next version of iPad OS would be needed for the initial release.

Of course, this has clearly gone off topic so to bring it back on topic the same would apply to Mac mini. I think it'll stay Intel.

If Apple are thinking of introducing an ARM powered box it's clearly got to be an AppleTV - and one with a lot of horsepower under the hood. I honestly can't see why Apple would want to allow people to hack up an ARM powered box though. That's a job for an (intel) Mac mini.

 

[Spectrum]

Isn't the AppleTV already an ARM chip?

If Apple are thinking of introducing an ARM powered box it's clearly got to be an AppleTV - and one with a lot of horsepower under the hood. I honestly can't see why Apple would want to allow people to hack up an ARM powered box though. That's a job for an (intel) Mac mini.

 

[sublunar]

Isn't the AppleTV already an ARM chip?

That was my point. Enhancing the already existing AppleTV.

 

[Spectrum]

That was my point. Enhancing the already existing AppleTV.

But they wouldn't be introducing an ARM-powered box then...they'd be making somethign that exists more powerful.

To me the most likely thing to go ARM (if anything is) is a MacBook ultraportable. Catalina is the start of this: encourage developers to tweak iOS apps to work well in a desktop environment - then simply run them natively on a new iOS ARM laptop.

 

[sublunar]

But they wouldn't be introducing an ARM-powered box then...they'd be making somethign that exists more powerful.

To me the most likely thing to go ARM (if anything is) is a MacBook ultraportable. Catalina is the start of this: encourage developers to tweak iOS apps to work well in a desktop environment - then simply run them natively on a new iOS ARM laptop.

I'd also likely agree with that but Apple have to be seriously looking at boosting the market for AppleTV+ by releasing new hardware and given that 2020 will be a year for games a '4k Pro' AppleTV would show off some of the better capabilities of the box. And yes, that still doesn't give hobbyists a cheap Mac mini that they are looking for. I wouldn't have thought Apple would be interested in making an ARM Mac mini for that reason and the Colo guys likely wouldn't be too interested either if the software isn't going to be there either - even if they could - again - recycle the existing form factor.

Not sure if I said it in this thread but I think an ultraportable ARM machine is even easier for Apple. Put a 12" iPad Pro in a laptop configuration with permanent scissor keyboard running iPad OS with USB-C port alongside the Lightning port.

I would say an A14X CPU would be ready for the big time if it's placed inside a device that isn't compromised with RAM and storage - for example, have minimum 8Gb RAM and 128Gb storage like the laptops. The Clear selling advantage here could be to claim 20 hours battery life, nice keyboard, cellular, full iOS support and USB-C/lightning connectivity.

This in effect makes it a replacement for the MacBook and there's already some big name apps available for iOS such as Office 365, Apple's own Pages, Numbers, and Keynote as well as plethora of iOS apps PLUS the ability to take a SIM card for ultimate connectivity - something that has been mooted for various MacBooks over the years.

Something more dramatic would be a fully haptic Touch Bar style keyboard (essentially making it a 2 screen device with a hinge) but that would make it exceptionally expensive rather than something like a $899-999 option. That sort of setup may be of interest for artists etc though.

And to briefly mention the Mini for a moment, I could see Apple going with an iBook (for the sake of argument over the name) as the Mac mini replacement for the switcher. Its an iPad in an official laptop configuration. What's not to like?

 

[Spectrum]

I'd also likely agree with that but Apple have to be seriously looking at boosting the market for AppleTV+ by releasing new hardware and given that 2020 will be a year for games a '4k Pro' AppleTV would show off some of the better capabilities of the box. And yes, that still doesn't give hobbyists a cheap Mac mini that they are looking for. I wouldn't have thought Apple would be interested in making an ARM Mac mini for that reason and the Colo guys likely wouldn't be too interested either if the software isn't going to be there either - even if they could - again - recycle the existing form factor.

Not sure if I said it in this thread but I think an ultraportable ARM machine is even easier for Apple. Put a 12" iPad Pro in a laptop configuration with permanent scissor keyboard running iPad OS with USB-C port alongside the Lightning port.

I would say an A14X CPU would be ready for the big time if it's placed inside a device that isn't compromised with RAM and storage - for example, have minimum 8Gb RAM and 128Gb storage like the laptops. The Clear selling advantage here could be to claim 20 hours battery life, nice keyboard, cellular, full iOS support and USB-C/lightning connectivity.

This in effect makes it a replacement for the MacBook and there's already some big name apps available for iOS such as Office 365, Apple's own Pages, Numbers, and Keynote as well as plethora of iOS apps PLUS the ability to take a SIM card for ultimate connectivity - something that has been mooted for various MacBooks over the years.

Something more dramatic would be a fully haptic Touch Bar style keyboard (essentially making it a 2 screen device with a hinge) but that would make it exceptionally expensive rather than something like a $899-999 option. That sort of setup may be of interest for artists etc though.

And to briefly mention the Mini for a moment, I could see Apple going with an iBook (for the sake of argument over the name) as the Mac mini replacement for the switcher. Its an iPad in an official laptop configuration. What's not to like?

I agree - I just hope they don't go with a haptic touchscreen...unless they have some AppleMagicTM that doesn't make it hideous to type on.

 

[Krevnik]

To me the most likely thing to go ARM (if anything is) is a MacBook ultraportable. Catalina is the start of this: encourage developers to tweak iOS apps to work well in a desktop environment - then simply run them natively on a new iOS ARM laptop.

I agree with the bolded bit, but not for the reasons you give.

Catalyst doesn't build ARM binaries. Just like the Simulator for iOS devices uses x64, Catalyst builds for x64. The main difference is that Catalyst brings UIKit and iOS-like entitlements to the Mac. That's pretty much it.

The ability to deploy bytecode to the App Store is what makes moving to ARM easier: by letting Apple compile down to machine code on their end when a new architecture is introduced.

 

[Spectrum]

I agree with the bolded bit, but not for the reasons you give.

Catalyst doesn't build ARM binaries. Just like the Simulator for iOS devices uses x64, Catalyst builds for x64. The main difference is that Catalyst brings UIKit and iOS-like entitlements to the Mac. That's pretty much it.

The ability to deploy bytecode to the App Store is what makes moving to ARM easier: by letting Apple compile down to machine code on their end when a new architecture is introduced.

Hmmm...my point was that the process of adopting Catalyst leads developers to optimise the UI, of what would otherwise be an iPad app, for a desktop use (unless I misunderstand things?). Thus the resulting UI-tweaked app can then be complied for iOS, and run on an ARM architecture natively.

Is that not correct?

 

[Krevnik]

Hmmm...my point was that the process of adopting Catalyst leads developers to optimise the UI, of what would otherwise be an iPad app, for a desktop use (unless I misunderstand things?). Thus the resulting UI-tweaked app can then be complied for iOS, and run on an ARM architecture natively.

Is that not correct?

Okay, I think I got this backwards, you're suggesting that iOS would be the laptop OS, not macOS.

iOS 13 certainly does provide a bunch of stuff that helps (that also enables Catalyst) and the iPad Pro should benefit from this as a side-effect to an extent, with more apps taking on keyboard shortcuts/etc. But other than multi-window support, iOS could already do a lot of this. Developers just were spotty in their support of it. Multi-window itself is and will be spotty for some time to come, sadly.

But I think Catalyst is more an answer to Electron and its ilk. One of the sad realities of macOS is the cases where popular applications are native on iOS, but some form of Web wrapper on Mac, taking up more space than the iOS app because of the middleware used. Trello is more than double the size on Mac, because it is built on Electron.

At this point, it'd be a lot easier to make an ARM macOS laptop than it would be to bring iOS up to the standard macOS sets in that space, IMO.

 

[rambo47]

When Apple announced the move to Intel processors, they also announced that they had been developing OSX for Intel from the start. Kept it in a secret development lab, paralleling the OSX-on-PPC development. The software was already (mostly) there, and porting apps to OSX-on-Intel was aided by Apple itself.

Any chance Apple has done this for OSX-on-ARM?

 

[Krevnik]

Any chance Apple has done this for OSX-on-ARM?

Outside the Window Server (and UIKit to talk to it), the included drivers, the default daemons, and file system layout, iOS *is* macOS on ARM.

So much code is already shared between the two at this point. It wouldn’t be zero work, but it isn’t starting from scratch either. The Window Server and AppKit would be the hardest parts, and jettisoning 32-bit support helps deal with a lot of legacy in the codebase (Carbon/etc.)

 

[dogslobber]

No new Mini for this year then. Guess Santa decided we'd not been good this year.