MacInTouch Amazon link...
Channels
Apple, Products, Questions


I can't help but think the T2 processor is primarily to stem the creation of hackintoshes. Throw some custom silicon in it, require it for basic functions like booting, and hackintoshes go bye-bye.
Or it could be about system security, which, frankly, I value quite highly. Having another processor with a different OS, designed with secure operation in mind, makes a lot of sense.

That's not saying you're not correct, but, frankly, just how much does the hackintosh market hurt Apple's sales?
 


Ric Ford

MacInTouch
The Geekbench guy did some developer workload (compile) tests to look at throttling of the 2018 MacBook Pro (but didn't have access to an i9 CPU to test):
John Poole/Primate Labs said:
MacBook Pro (Mid 2018) Throttling
...
Conclusions

So what’s going on here? Why does this test not replicate the throttling seen in other tests? Part of the issue is the tests themselves. Premiere uses both the CPU and the GPU, while Geekbench only uses the CPU. If the GPU contributes significant heat, then that will cause the CPU to throttle more aggressively. It’s possible the decrease in performance observed in Premiere is due to a combination of new AMD GPUs with new Intel processors, or to the new AMD GPUs themselves.

My recommendation? If your work doesn’t involve long-running tasks that are CPU- and GPU-intensive (such as Premiere) then the new MacBook Pro should provide a considerable increase in performance. Otherwise, it might be wise to wait until more performance data is available.
 


I presume that these new MacBook Pro models won't be able to run Sierra. (I'm still leery about High Sierra and APFS.)
 


Ric Ford

MacInTouch
Here's a fun software toy for CPU testing:
Intel said:
Intel Power Gadget

Intel® Power Gadget is a software-based power usage monitoring tool enabled for Intel® Core™ processors (from 2nd Generation up to 7th Generation Intel® Core™ processors). Intel® Atom™ processors are not supported. It is supported on Windows* and Mac OS X* and includes an application, driver, and libraries to monitor and estimate real-time processor package power information in watts using the energy counters in the processor.

... Traditional methods to estimate power/energy usage of the processor has always been a cumbersome task that included special purpose tools or instrumentation on the platform along with third party equipment. The motivation for the tool was to assist end-users, ISV’s, OEM’s, developers, and others interested in a more precise estimation of power from a software level without any H/W instrumentation.
 


Ric Ford

MacInTouch
Still more on the Touch Bar MacBook Pro keyboard problems and changes:
iFixit said:
Anatomy of a Butterfly (Keyboard)—Teardown Style

... Since Apple informed their service providers that the membrane is “to prevent debris,” we’re inclined to think any change in noise level really is just a secondary feature.

... Apple has a proven track record of failure for these keyboards. They’re being accused, by way of several class-action lawsuits, of knowingly selling failure-prone keyboards. Apple may claim that they design products to last—and that designing for repairability compromises the durability of a device—but this keyboard misadventure belies those points. If a single grain of sand can bring a computer to a grinding halt, that’s not built to last. If said computer can only be fixed by throwing half of it away and starting over, that’s not built to last. We’re definitely excited to see improved protection on these machines—consumers deserve it with the prices they’re paying.
 


Another analysis of the throttling issues with the 2018 MacBook Pro:
ItsQuadPod/Reddit said:
What's In a TDP? Examining the Thermal Issues in the New MacBook Pro

While the old CPUs could only produce up to 45W at full boost clocks, the Coffee Lake chips will now draw up to 100W at their own. This amount of power is far beyond what the MacBook Pro was designed for, and will only reasonably be held for a short amount of time before the temperature spikes way up, at which point the CPU should throttle back down to its base clock speed where it produces 45W and the cooling system can keep the clock speed steady. However, the fans in Macs take notoriously long to ramp up so this does not happen and clock speeds tend to rapidly spike up and down without much control right now. Additionally, in loads where the CPU and GPU must work at the same time clock speeds fall even more than before thanks to those two extra cores and performance may drop below the old levels.
This makes it sound like Apple might be able to do some tuning on the fan speeds to at least partially mitigate the problem.

My impression (based on what information is out right now) is that the keyboard issues should (mostly? completely?) be mitigated with the membrane, and if your use case is such that you'll be working both the CPU and GPU hard on a regular basis, you're better off with an iMac or Mac Pro. Perhaps an external GPU might help mitigate this on the MacBook Pro, but you lose laptop portability at that point.
 


I can't help but think the T2 processor is primarily to stem the creation of hackintoshes. Throw some custom silicon in it, require it for basic functions like booting, and hackintoshes go bye-bye.
I might get that paranoid if the rumored (yet again) upgraded Mini has a T2 chip, but I'm guessing it won't. Aside from a couple of egregious cases of people hawking systems with Apple components, Apple has pretty much left the clone/Hackinstosh market alone since August, 1997, when they bought out Power Computing. I think the management at Apple is sensible enough to realize that those people willing to literally hack around to build a Hackintosh don't represent much in terms of lost sales to Apple — most would just build Windows or Linux machines if their access to the OS were cut off.

A more serious discussion topic might be whether the average user needs or wants the level of security that data-at-rest encryption provides, at the cost of time-consuming and even potentially disastrous experiences if hardware damage occurs. I suppose Apple knows the internal component IDs used for hashes in T2-equipped models and could custom-burn a new T2 for owners of record who'd had theirs destroyed somehow... but somehow I doubt that's ever going to happen.
 


I can't help but think the T2 processor is primarily to stem the creation of hackintoshes. Throw some custom silicon in it, require it for basic functions like booting, and hackintoshes go bye-bye.
I doubt it. In order to make that work, Apple would have to modify macOS to refuse to boot on any Mac without a T2 chip. Maybe that will be possible in a few years, but right now most Macs (all but the newest MacBook Pro and iMac Pro) don't have the chip. Apple isn't going to make an OS change that breaks their entire user base. And even if they wanted to, could you imagine the lawsuits that would ensue if the next macOS couldn't boot on any Mac other than the two newest models, including ones still in production?

Keep in mind also that people were posting similar fears when the first Intel Macs were announced, because those first prototype Intel Mac developer workstations all had TPM chips on the motherboard. Everybody was afraid that Apple was going to lock the OS to TPM-hosted credentials. It didn't happen. Quite the opposite, the Intel Macs that actually shipped didn't even have TPM chips.

In other words, while it is theoretically possible that Apple might do what you're afraid of, I wouldn't lose any sleep over it. They've had this ability for a very long time and have not decided to do anything so far. The presence of the T2 chip without any additional facts doesn't sound like a strong enough argument to conclude that they're changing their opinion after all these years.

Of course, if I'm wrong and Apple does decide to lock down macOS like this, I think you can rest assured that the Hackintosh community will find the relevant parts of the OS and disable and/or replace them. The iOS Jailbreak community has proven that there is no such thing as a lock that can't be picked.
 


I can't help but think the T2 processor is primarily to stem the creation of hackintoshes. Throw some custom silicon in it, require it for basic functions like booting, and hackintoshes go bye-bye.
I genuinely can't imagine that Apple cares that much. The number of sales lost to Hackintoshes has got to be a rounding error to Apple; the absolute number can't be that large, and some significant fraction of those who currently build them probably wouldn't be buying a new Apple machine if they couldn't. Given the engineering cost in designing the T2 and its subsystem, building support for it into the OS, and eventually including a chip in every Mac the company ships, I'd be shocked if that could possibly make financial sense if its goal was the elimination of Hackintoshes.

On the other hand, it's doing some very cool stuff, and may enable even more in the future by virtue of being a fairly powerful CPU with its own micro-OS. I don't think anyone knows clock speed on it, but with a single A10 comparable core, it could well be as fast as one of the cores in the 2016 MacBook (or one core of a 2011 or 2012-era i5 laptop CPU), which is really impressive if you think about it.

As a privacy advocate I'm quite pleased with the cryptography that Apple has gone all-in on. I likewise doubt that Apple's aggressive security is enough of a sales point to make the engineering on it break even from a financial perspective, since consumers mostly just don't get or care about security, so I appreciate its inclusion all the more.
 


Ric Ford

MacInTouch
I had forgotten about this hidden limitation in the 2016-2017 Touch Bar MacBook Pros that is resolved in the 2018 redesign:
The Verge said:
Apple’s new 13-inch Touch Bar MacBook Pros now have four full-speed Thunderbolt 3 ports
It’s a small but useful update to the earlier 2016 and 2017 13-inch models, which only offered full-speed data on the two left-side ports, while the two right-side ports were limited to slower PCI Express bandwidth. The reason for the shift is simple: the original Intel chips on the old 13-inch MacBooks only offered 12 PCI Express lanes, which wasn’t enough for the full 40Gbps transfer speeds on all four ports.

The new eighth-generation chips on the 13-inch model now offer up to 16 PCI Express lanes, which is enough for all four ports to operate at full speed, just like its larger, 15-inch sibling.
 


I can't help but think the T2 processor is primarily to stem the creation of hackintoshes. Throw some custom silicon in it, require it for basic functions like booting, and hackintoshes go bye-bye.
It's possible, obviously, but will take a long time for this hardware to filter through the entire product line. So it seems several years off - at minimum - before this could become a requirement to run macOS.

More realistically, hackintoshes simply won't run features that are important for security (a limitation that currently applies to FileVault on boot drives).

And I doubt that Apple is particularly concerned about hackintoshes as a hit to their market - it's marginal.

I think the T2 should be seen as a trade-off for customers (business, government) that have stringent security requirements - for whom the added security of the T2 chip integrated with the OS may be a significant benefit. Or to formulate as a question, does anyone know of any existing or pending security guidelines or standards that may require the type of lock-down a T2-type chip can enable? I wouldn't be surprised if such requirements drive a wave of upgrades in future, to Apple's benefit.
 


Another analysis of the throttling issues with the 2018 MacBook Pro:
This makes it sound like Apple might be able to do some tuning on the fan speeds to at least partially mitigate the problem....
Given the information out there now, that sounds like a possibility.

One thing I've been using on my 2012 15" is TG Pro (from Tunabelly), since it allows you to control when and how quickly the fans come on. It helps to keep the Mac cooler than it would be on its own. I do hear the fans more often, but since the 2011 17" died, I've been using it to (hopefully) avoid thermal damage.

I've been considering replacing the 2012 15" with the 2018 myself, but the thermal issues gave me pause. The new information further up in this thread is encouraging.

(No affiliation with TG Pro, by the way, just someone who has used it for 4 years now.)
 


I am now wondering about the thermal issues with the i9 chip, since we are now seeing more reports and not depending on a single data set. I do not think tweaking the fans is going to help that much, since some testing was done with the fans at full power.

I am looking forward to a battery of tests being run on all 13" and 15" systems. My suspicion is that the issue is limited to the 15" i9, but we need solid, consistent tests. We should also see what triggers the problem. Is it only when all 6 cores are at full speed? I think the Final Cut Pro and Premiere tests are the most useful, since they represent actual use of many people who are interested in the i9 configuration, and they run the CPU at full speed.
 


I might get that paranoid if the rumored (yet again) upgraded Mini has a T2 chip, but I'm guessing it won't.
The T2 chip is, among a few other things, Apple's default SSD for Macs. (Well, technically, it is the SSD controller, but with Apple's embedded and flash NAND only board designs, pragmatically it is the SSD, as far as the system is concerned).

Apple's march to making all Macs primarily SSD-based should be relatively transparent at this point. The only hold-up is SSDs being "affordable enough" for the entry-level models. There are no Mac laptops with a hard disk drive at this point, and the top end of the desktop line-up is SSD only.

Apple wants to put SSDs in all Macs. T2 is an SSD. Therefore, don't be surprised when some later variant of the T2 appears in all Macs somewhere down the road. If flash memory cost the same $/GB as rotating platter storage, Apple would have already made the move across the board. The T2 will keep Apple's SSD prices a bit higher than mainstream.

I think it is doubtful that Apple will shift the T2 boot policies so that the singular internal Apple SSD is the only possible drive the Mac can boot off. As long as the Macs can boot off of non-Apple SSDs, then the path to making a non-Apple system work is relatively viable.

Thunderbolt pragmatically allows a path for low-level boot drives to be added to a Mac system, so the concept of additional drives other than the singular Apple boot drive isn't going away any time soon. I think Apple is going to optimize along the default, "out of the box" experience and make it a secure boot with just Apple stuff, but they aren't going to kill off the additional value-add of expanding the system in all usages. (That isn't internal components, that is any possible expansion.)

I think the management at Apple is sensible enough to realize that those people willing to literally hack around to build a Hackintosh don't represent much in terms of lost sales to Apple — most would just build Windows or Linux machines if their access to the OS were cut off.
I'd guess the percentage is down in the 1% range. That stays manageable as long as Mac sales are growing (even at a slow rate close to flat). If Mac sales fall backwards significantly, that would be a problem.

It is unlikely most of those folks won't largely 'shift' to Linux/Windows. A bigger 'war' of measure-vs-counter-measure would likely escalate.
A more serious discussion topic might be whether the average user needs or wants the level of security that data-at-rest encryption provides, at the cost of time-consuming and even potentially disastrous experiences if hardware damage occurs.
Hardware damage isn't the most likely issue. The majority of users have mobile computers. Those get lost, stolen, compromised, etc. at a far higher rate than some drive hardware failure.

The "time-consuming" cost if you have a back-up is a restore (which is a couple of hours and close to free in cost). That isn't that high.

As SSDs get in the zone of triple- and quad-level charge storage in a single cell, the low-evel recovery process is going to get harder as the schemes to encode and store the data get more complicated. If the SSD controller gets borked and the metadata gets fried, there isn't going to be a whole lot of recovery, even if that data is not encrypted.

Modern SSDs don't do 1-to-1 block mapping from logical address to physical. Even without encryption, all the data is encoded in some way. The notion of "I'll get a magnet and read it straight off the media" to recover the data overlooks what is actually under the covers in modern SSDs.

I suppose Apple knows the internal component IDs used for hashes in T2-equipped models and could custom-burn a new T2 for owners of record who'd had theirs destroyed somehow... but somehow I doubt that's ever going to happen.
I doubt they have a record. If they did, it would be subject to a 'subpoena' in whatever country they operated in. Also eliminates the chance a rogue employee will steal the keys and significantly damage their security reputation. If not there, it can't be stolen.
 


Apple’s new 13-inch Touch Bar MacBook Pros now have four full-speed Thunderbolt 3 ports
The reason for the shift is simple: the original Intel chips on the old 13-inch MacBooks only offered 12 PCI Express lanes, which wasn’t enough for the full 40Gbps transfer speeds on all four ports.
First, Thunderbolt is 40Gbp/s from the Thunderbolt controllers. That is the Thunderbolt speed; not the PCI-e speed. Second, x4 PCI-e v3 is 32Gb/s. Thunderbolt does 40, so that can deliver 32Gb/s in a timely, relatively low-latency fashion. If delivering other protocols, you need to be faster than what you are hauling/delivering.

A Thunderbolt controller only takes an x4 PCI-e v3 connection. Each controller does two ports, so you only need two controllers for a 4-port system. Two Thunderbolt controllers just need x8 lanes. 12 > 8, so this explanation is more than a bit odd. There is no discrete GPU in the 13" model, so not sure what would be consuming the PCI links off the CPU. The rest of the I/O would be hanging of the PCH chip. I suspect it is not the number, but the default grouping the CPU allows (but that's odd, too, because 3 by 4x would be a natural way to group 12. The only problem would be if it was x4 and then couldn't chunk another x4 out of the remaining 8).

I'd suspect a board layout problem and/or costs as being the problem far more than lack of gross oversupply of PCI-e lanes. For example, if Apple used an x2 connection to the Thunderbolt controller, there would be lower PCI-e bandwidth to that controller. That doesn't mean the controller couldn't approach 40GB/s if fed with two DisplayPort feeds plus that x2 feed.
which is enough for all four ports to operate at full speed, just like its larger, 15-inch sibling.
Actually, it is the 15" model with a discrete GPU than has the larger bandwidth restriction. The GPU is soaking up x8 (probably not all x16) of the PCI-e lanes coming out of the CPU package. The lack of a dGPU in the 13" model means there are more unused bandwidth, not less.
 


I am now wondering about the thermal issues with the i9 chip...
The root issue isn't the i9 chip package, it is the overall system. The natural Apple Store selection path to an i9 MacBook Pro 15" bundles it with the highest GPU also. (There is another BTO path by selecting the lowest-price system and bumping up the CPU but not the GPU.)

A 555X versus 560X probably would completely solve the problem, but should provide a small bit of "headroom" inside the system. If the heat pipe design stayed the same, the heat pipes for the CPU and GPU are coupled, to provide each with outflow access to each of the two fans, but on max load from both, that should be a coupled throttling issue on both.
We should also see what triggers the problem. Is it only when all 6 cores are at full speed? I think the Final Cut Pro and Premiere tests are the most useful, since they represent actual use of many people who are interested in the i9 configuration, and they run the CPU at full speed.
Concurrent max CPU and GPU is quite likely a trigger. The other two 15" CPU offerings - i7-8850H and i7-8750H - have TDP configurable down to 35W (Apple apparently is not running in that mode, but they can be stepped down). The i9 model can't. It is probably just barely squeezing under the 45W TDP limit. If you throw in a GPU upgrade, which also has some TDP scope creep, and don't change the physical cooling system parameters at all, then this result wouldn't be surprising at all.

Most of the rest of the configs would work OK, but max CPU + max GPU isn't the focus of the design. Apple dropped the old 15" design, but if Apple had a "desktop replacement" oriented model also in the line up, that would have been a more natural fit for the i9 plus as big (hot) of a GPU as they could fit.
 


I genuinely can't imagine that Apple cares that much. The number of sales lost to Hackintoshes has got to be a rounding error to Apple; the absolute number can't be that large, and some significant fraction of those who currently build them probably wouldn't be buying a new Apple machine if they couldn't.
Not just that, but that small number of Hacintosh users are socially well connected, often Apple customers with another device, or possibly future Apple customers. Alienating them by stomping their small, active community would likely snowball and hurt Apple way more than ignoring it would.
 



Not just that, but that small number of Hacintosh users are socially well connected, often Apple customers with another device, or possibly future Apple customers. Alienating them by stomping their small, active community would likely snowball and hurt Apple way more than ignoring it would.
I built a hackintosh on a self-built homebrew machine similar to a pro model and installed Mac OS X 10.6 through about 10.9. It was a lot of bother and a good time-waster as opposed to other time-wasters.

We own eight operable Macs, as it is. The most efficient way to be a hackintosh owner/builder is to own a Mac, so you can easily download the OS installer. So, I doubt Apple has really lost much business.
 


The Geekbench guy did some developer workload (compile) tests to look at throttling of the 2018 MacBook Pro (but didn't have access to an i9 CPU to test):
One time I was using Handbrake on my older MacBook Pro. Handbrake has a realtime readout of how fast it is going in frames per second. Handbrake pinned all cores (4) to 100% with the fans running on full. Being an engineer, I was curious about throttling.

Being winter and about 20 degrees F outside and having an outside table next to a receptacle, I made a test. I took a metal oven pan that, when turned upside down, was not as wide as the feet on the bottom of the MacBook Pro - that is, the bottom of the MacBook Pro sat directly on the oven pan. Thus, I had a large metal heat sink operating with an ambient temperature of 20 degrees F.

While it was running Handbrake, I took the MacBook Pro outside, set it directly on the metal oven pan, and plugged it in. In about 5 minutes, the speed of Handbrake went up about 10 to 15%, and the fans went off.
 


I can't help but think the T2 processor is primarily to stem the creation of hackintoshes.
Not sure what Apple's plans are for the T2 but hackintosh elimination seems unlikely. Folks looking for potentially nefarious schemes might consider Gatekeeper and the new notary service. Apple generates significant income collecting 30% from developers and the revenue stream keeps increasing.
 


Not just that, but that small number of Hacintosh users are socially well connected, often Apple customers with another device, or possibly future Apple customers. Alienating them by stomping their small, active community would likely snowball and hurt Apple way more than ignoring it would.
Not p**sing off the 80+ million current Mac users is a far higher priority for Apple than the social networks of a relatively small handful of folks. Apple has a large network of stores with millions of visitors per month. Apple socially interacts with millions of people per day. Apple has a social network that is far bigger than the edgy fringe. Apple is way past the stage where they need "guerrilla", evangelist marketing to sell product. Apple actually running explicit ads for Macs only adds to that.

Trying to completely shut off Hackintosh as close to 100% as possible would likely blow back into the deployed, older Mac hardware network. That blowback isn't worth it.

When UEFI pragmatically finally dumps BIOS support, then I could see Apple perhaps merging their firmware closer to UEFI and then perhaps needing some additional differentiation on boot, to moderate Hackintosh growth, but that would likely only happen over an extended period of time (as the old non-T2-like Macs retire out of support windows). If Apple could dump extra Windows driver work and simplify multiple OS booting, that would be a likely course. T2 also greases the long-term path to an 'exit' to Mac ARM systems. I don't think they want to do that at the moment, but it doesn't hurt them to enable the option.

The longer it takes Apple to put 'T2' boot hardware into the whole Mac line up (plus 7 years) the longer window to even plausibly squeezing out Hackintosh completely. That is so far out that technology will be substantively different and the market drivers changed from the current ones, also.

If Apple did a 90-degree left turn into a 100% line up of ARM-primary CPU Macs, that would kill off Hackintoshes far more than what the T2 feature set is doing. (Diverging from the shared sub-components base of the Windows hardware market would be a bigger issue.) There is exceedingly little to indicate that is the radical path Apple is on, though. Adding a ARM chip in a supplementary role doesn't mean trying to put it into the primary role across the board.
 


I am now wondering about the thermal issues with the i9 chip, since we are now seeing more reports and not depending on a single data set. I do not think tweaking the fans is going to help that much, since some testing was done with the fans at full power.
I wonder how much better it might do with some kind of active cooling base. There are many laptop stands with integrated fans and I think I've seen a few with Peltier coolers as well. I'd love to see if the benchmarks change when the lower case is kept chilled by such a stand.
 


A potential thermal mitigation (not a solution) is a software fan control, such as Macs Fan Control. It allows you to set a temperature range and increases/decreases fan speed accordingly.

My iMac is set so the fans start running much earlier (i.e. starting at 89˚ F with a top range of 113˚ F and using the CPU Proximity sensor). I'd like to believe the long life of my own Macs is partially attributable (but have no demonstrable evidence) to running a little cooler. I'm not affiliated in any way with this software vendor.
 


The root issue isn't the i9 chip package, it is the overall system. The natural Apple Store selection path to an i9 MacBook Pro 15" bundles it with the highest GPU also. (There is another BTO path by selecting the lowest-price system and bumping up the CPU but not the GPU.)

A 555X versus 560X probably would completely solve the problem, but should provide a small bit of "headroom" inside the system. If the heat pipe design stayed the same, the heat pipes for the CPU and GPU are coupled, to provide each with outflow access to each of the two fans, but on max load from both, that should be a coupled throttling issue on both.

Concurrent max CPU and GPU is quite likely a trigger. The other two 15" CPU offerings - i7-8850H and i7-8750H - have TDP configurable down to 35W (Apple apparently is not running in that mode, but they can be stepped down). The i9 model can't. It is probably just barely squeezing under the 45W TDP limit. If you throw in a GPU upgrade, which also has some TDP scope creep, and don't change the physical cooling system parameters at all, then this result wouldn't be surprising at all.

Most of the rest of the configs would work OK, but max CPU + max GPU isn't the focus of the design. Apple dropped the old 15" design, but if Apple had a "desktop replacement" oriented model also in the line up, that would have been a more natural fit for the i9 plus as big (hot) of a GPU as they could fit.
This always annoyed me with Apple's cattlechute options: I see a portable running an i5 with fastest GPU as practical for me in battery and performance. (Or even one notch down in GPU option.) Even an i7 with a GPU just under the top option works for me with production.

But they put an i9 (desktop processor) in a laptop... what were they thinking? Who is their target for this?

(Oh, I would like an i9 Mac Mini with 16 GB and a 1TB SSD with 555X GPU... but that would sap iMac Pro sales.... ;)
 


Lisa at MobileTechReview has posted "sensible" reviews of several recent "thinner and lighter" laptops, noting that "thinner" computers have more thermal issues than "thicker" ones with more space for air flow.

She just posted her review of the new MacBook Pro, comparing it to a thicker Razer gaming laptop, and noting differences between Adobe Premiere and Final Cut on Mac and benefits of controlling Mac fan speed.

MobileTechReview/YouTube said:
 


I haven't completely disassembled a recent MacBook Pro. Can anyone verify if the newer keyboards are actually glued in?
And a few days later, iFixit did a teardown of the new keyboard, and it's definitely not serviceable.

To start, the underside of the keyboard is covered with a glued-on shield. Then in order to disconnect the keyboard from the upper-case, there are lots of P2 pentalobe screws and a bunch of rivets that need to be cut away. And then more glue.

Oh yes, and if you don't want to tear that new silicone membrane, you may need to remove all the keycaps from the board before you do the above procedure. iFixit did this. Looking at the photos, the membrane might come loose with the keyboard if the caps are left in place, but that's very unclear - it it's stuck (glued?) to the aluminum, then those caps are going to have to come off to avoid ripping it.

So I guess it's theoretically repairable, but I doubt anybody will be crazy enough to try. If the keyboard dies after the warranty expires, I recommend just switching to an external keyboard. Anything else will be either too expensive (replace entire top-case assembly) or too difficult (replace keyboard).

Sadly, iFixit's deep dive into the new keyboard shows that the silicone barrier helps a little, but sand can still get in there and ruin it. So it's still no good, in my opinion.
 


I presume that these new MacBook Pro models won't be able to run Sierra. (I'm still leery about High Sierra and APFS.)
Nope. Apple [almost] never allows an OS older than the [latest] to run on new machines. It is possible to install High Sierra on a Journaled HFS+ formatted SSD drive, but I am not sure that it can done with the new 2018 MacBook Pros. You need a full installer and to create a bootable USB drive. I don't think Apple has released a full installer that supports the 2018 MacBook Pros - the version of macOS 10.13.6 on the App Store is the wrong build.
 


I might get that paranoid if the rumored (yet again) upgraded Mini has a T2 chip, but I'm guessing it won't..
I think it is almost a certainty that all new Macs will have the T2 (or its successor) installed. Apple is using the T2 chip to enforce security and encryption. Apple very rarely reverses course once they start down the road.

But, I don't see that hampering hackintoshes. It will be many years before Apple can enforce the existence of a certain chip to allow boot.
 


Ric Ford

MacInTouch
Nope. Apple [almost] never allows an OS older than the [latest] to run on new machines. It is possible to install High Sierra on a Journaled HFS+ formatted SSD drive, but I am not sure that it can done with the new 2018 MacBook Pros. You need a full installer and to create a bootable USB drive. I don't think Apple has released a full installer that supports the 2018 MacBook Pros - the version of macOS 10.13.6 on the App Store is the wrong build.
The key here is that new Mac hardware typically requires new driver software at a minimum to support that new hardware, and that special hardware support is typically only added to the very latest macOS build, not added back to earlier versions. In fact, new models often require special new versions of the OS that have never been released before, containing hardware drivers that may never be back-ported to earlier versions.

This can be very confusing because of issues with Apple's version numbers. In the case of the 2018 MacBook Pro, their new 8th-generation CPUs and other hardware changes require exactly this type of special new macOS build. Apple's specs are frustratingly vague, but the EveryMac database has more useful information:
everymac.com said:
The build number (17G2112) is a key here, differentiating among different "10.13.6" versions which do or don't support the new hardware, as described here:
tonymacx86 said:
Native Intel UHD630 Graphics support in macOS 10.13.6

So the new 2018 MacBook Pros are finally here, and they use the new 8th Gen CPUs & Graphics!

Installing the Updated OS
While Apple hasn't released a new macOS installer (current is build 17G65, while 17G2112 has support for new laptops), we can still get it directly from Apple's servers. macOS Recovery is using the same method to download macOS, so we can use it to get the latest available build....
 


... But they put an i9 (desktop processor) in a laptop... what were they thinking? Who is their target for this? (Oh, I would like an i9 Mac Mini with 16 GB and a 1TB SSD with 555X GPU... but that would sap iMac Pro sales.... ;)
The i9 in the MacBook Pro is substantially different from the desktop-class i9 implementations. The laptop variant that Apple is using is probably the same silicon die as the other [laptop-class] ones being used.

"entry" 2018 MacBook Pro 15":
i7 8750H
base 2.2GHz, turbo 4.1GHz

"top" 2018 MacBook Pro 15"
i9 8750HK
base 2.90GHz, turbo 4.80GHz

The bulk of the rest of the specs are the same. There is a L3 cache size difference, but that may not be a die difference. There is a "thermal velocity boost", which probably doesn't do much in the MacBook Pro but is probably not a die difference, either.

Same number of cores. Same integrated GPU. Same memory specs. Same PCI-e specs. All suggestive of dealing with exactly the same silicon die here. It is no more a desktop chip than the vast bulk of the other i5 and i7 "laptop" chips. Thy are all the same baseline design. There are some GPU tweaks on some but for most part it is largely the same die with features turned on/off for marketing (and binned for performance).

Intel's mainstream 8th-gen desktop chips run in the 60-95W range. (Clocks are up and GPUs mostly smaller, but same baseline reference design as these 'laptop' implementations.) This laptop i9 has the same TDP as the Core i7-7820HK Apple used back in 2017 models. The i9 just runs right up to the line as the TDP range allows.

If Apple had designed the MacBook Pro in the previous generation with some reasonable buffer just above 45W, this wouldn't be a problem. Instead, Apple apparently measures right to the component they are using and shrinks the envelope to almost exactly just enough not to be a problem. (Rinse and repeat the same practice on the GPU... which probably is running over TDP if pushed hard, since AMD is largely just cranking the clocks for these speed-bumped versions.)

Intel's Core M (Y-series) are a different baseline design than the laptop/desktop series.
 


The key here is that new Mac hardware typically requires new driver software at a minimum to support that new hardware, and that special hardware support is typically only added to the very latest macOS build, not added back to earlier versions. In fact, new models often require special new versions of the OS that have never been released before, containing hardware drivers that may never be back-ported to earlier versions.

This can be very confusing because of issues with Apple's version numbers. In the case of the 2018 MacBook Pro, their new 8th-generation CPUs and other hardware changes require exactly this type of special new macOS build. Apple's specs are frustratingly vague, but the EveryMac database has more useful information:

The build number (17G2112) is a key here, differentiating among different "10.13.6" versions which do or don't support the new hardware, as described here:
If someone wants/needs a copy of the the full installer for the MacBook Pro 2018, you can use this utility to download the installer:

https://github.com/munki/macadmin-scripts/blob/master/installinstallmacos.py

It will list all current versions of macOS (including betas) and allow a user to download it.
 


Ric Ford

MacInTouch
The 2018 MacBook Pros have some graphics limitations:
MacRumors said:
2018 MacBook Pros Have New 'Titan Ridge' Thunderbolt 3 Controller, But DisplayPort 1.4 Isn't Truly Supported
...
  • The new 15-inch MacBook Pro theoretically supports DisplayPort 1.4, which Apple confirmed, but at least for now, it still can't drive an 8K display. It could be possible with VESA's lossless Display Stream Compression standard, perhaps, but it's unclear if this can be enabled down the road.
  • For now, then, the new 13-inch and 15-inch models have the same compatibility with external displays as the previous-generation MacBook Pro: up to two 5K displays or up to four 4K displays on the 15-inch model, and up to one 5K display or up to two 4K displays on the 13-inch model.
For comparison, 2016 and 2017 MacBook Pro models are equipped with Intel's JHL6540 Thunderbolt 3 controller, which supports DisplayPort 1.2.
 


Ric Ford

MacInTouch
It looks like Dell's XPS 15 is having thermal throttling issues with the Core i9 + GPU also:
Reddit said:
XPS 15 i9...
As everyone had mentioned the cooling solution for the form factor of the XPS 15 is insufficient for the amount of heat that the laptop creates when gaming. You will experience some sort of throttling. Due to CPU throttling you won’t be utilizing the 1050ti fully as designed. When doing CPU benchmarks alone I was able to keep at least the baseline 2.9ghz for the i9 but it was really pushing the temps.

... Sorry for the rant but I spent the last few weeks really wanting to keep this laptop. In every respect it is the best PC laptop on the market but Dell made the mistake of putting in an i9 into a thermal package that is not suitable for it. You can try to use an external GPU to reduce the graphics load on the laptop but the fans will spin at full speed even if you put moderate load on it.
 


Ric Ford

MacInTouch
Here's a pretty good analysis of the new MacBook Pro thermal issues, expanding on some of the points Lyman made above:
SlashGear said:
Why Apple may be the victim in 2018 MacBook Pro throttling
... Big cooling systems can handle bigger TDPs, but also result in larger, less portable notebooks. Slimmer designs, meanwhile, could struggle with those higher TDPs. One prominent suggestion about the new MacBook Pro, therefore, is that its cooling provisions simply weren’t up to the challenge of the highest-performance the Core i9 is capable of. Teardowns of the new notebook seem to show that the design is pretty much the same as for the older, slower versions.

The blame can’t all be placed at Apple’s feet, though, at least according to the latest theories. The design of the current MacBook Pro was first revealed in late 2016, but will have been in development for several years before that. Apple will have worked with Intel on exactly how much cooling was required in order for its laptops to work at peak performance.

That, though, depends on Intel’s roadmap for its processors, and it’s there where some have suggested things have struggled. Intel has had some widely-reported challenges getting its chips down to the sizes it initially promised, with its transition to 10nm processes taking far longer than it initially hoped for. That also means that the TDPs it was aiming to achieve – and the TDPs with which Apple undoubtedly designed the MacBook Pro’s cooling in mind – simply haven’t been achieved.

As a result, the Core i9-8950HK which Apple is offering in the 2018 MacBook Pro has a TDP of 45 watts. That, however, is only when it’s running at its regular clock speed of 2.9 GHz. Pushed further, to the Turbo Boost clock of 4.8 GHz, and you’ll be getting significantly more heat.
 


Ric Ford

MacInTouch
This YouTuber does a pretty extensive set of tests comparing the 2018 MacBook Pro i9 model with the low-end i7 model across a variety of apps and data, including Premiere, Final Cut and Resolve video processing. The high-end i9 was faster in all tests vs. the low-end i7, while different apps and data created radically different conditions and results.
Jonathan Morrison/YouTube said:
Everyone is Wrong About the i9 MacBook Pro (2018)
... I think this is gonna end up being a multi-part series. In this first video I'm gonna tackle the i9 versus the i7 and kind of address the issue of, if the more expensive i9 core is actually slower than the cheaper six-core i7, and spoiler alert, it's not. Now before I jump into my test, just to quickly explain why I think Dave's video was a little skewed is the test he used was honestly the perfect recipe to completely cripple your computer - he took 5K RED raw footage, which will eat your computer alive, brought that into Premiere Pro, which isn't optimized to begin with and then exported that into a 4K H.264 file.
 


Being winter and about 20 degrees F outside and having an outside table next to a receptacle, I made a test. I took a metal oven pan that, when turned upside down, was not as wide as the feet on the bottom of the MacBook Pro - that is, the bottom of the MacBook Pro sat directly on the oven pan. Thus, I had a large metal heat sink operating with an ambient temperature of 20 degrees F.
I have for a number of years used an old-fashioned wire in-basket as a stand for my MacBook Pro. My old 2011 17" models fit great; the current 15" models are just barely large enough to rest on the top of the basket. There are two benefits I get from doing this: air circulation under the laptop, and the bottom of the laptop screen is almost perfectly aligned with the bottom of my Apple Cinema 30" displays (I have one at work and one at home).
 



I think the T2 should be seen as a trade-off for customers (business, government) that have stringent security requirements - for whom the added security of the T2 chip integrated with the OS may be a significant benefit. Or to formulate as a question, does anyone know of any existing or pending security guidelines or standards that may require the type of lock-down a T2-type chip can enable? I wouldn't be surprised if such requirements drive a wave of upgrades in future, to Apple's benefit.
Let's say, hypothetically, that I work for such an outfit. Let's suppose, equally hypothetically, that my outfit has a requirement for data-at-rest encryption for all portable devices, as well as any desktops whose volumes contain any material referred to as "sensitive but unclassified." Then my employer might be interested such a solution. Several years ago, our then lead sys admin attended a talk by a major SSD manufacturer. The lecturer said they were all in with encryption on the PCB, settable by SATA command, to the nods and smiles of the management types from our organization present. I don't know whether encryption is available directly on as small form-factor NVMe storage as Apple's using in the iMac Pro or in the bare chips in the 2018 MacBook Pro, but I guess the T2 is performing that function instead. Given Apple's well-publicized stand on user privacy, certain outfits might well, again hypothetically, think it was A Good Idea. They might even enshrine it in our standards for hardware purchases someday, once Windows machines catch up.
 


Amazon disclaimer:
As an Amazon Associate I earn from qualifying purchases.

Latest posts