MacBook Pro with i9 chip is throttled due to thermal issues, claims YouTuber
macrumors.com
macrumors.com
His argument is such a slam dunk that the article concludes with stunned disbelief, speculating that maybe there's "something wrong with the MacBook Pro with Core i9 chip that Lee received".
Yes there's something wrong with it, that was the exact point of his video. Do people really think he just got a lemon?
If you watch a lot of his videos, you'll notice that he almost always talks about thermals, so it's not surprising to see him focus on that aspect of the MBP.
That, along with performance that's below the previous-gen i7 but charging $300 more is the real kicker - honestly, I'm not sure how anyone at Apple would be able to justify that as not just bilking their customers out of money.
It seems like these days, the mantra is "buyer beware" if you're looking to buy a Mac.
For audio, it's more the CPU.
This entire thread is baffling to me. Everyone rushing to justify buying a new high performance laptop which they don’t mind throttling because they never needed the performance in the first place??
Time scale is relevant. I want interactive tasks to be faster. I won't notice a 5% improvement in a job that takes long enough for a coffee break, but I will appreciate if more short tasks can cross the threshold of being indistinguishable from instantaneous.
For typical loads, both will be at a high level of usage
If what you care about is performance on 2-hour tasks, there's just no chance, that's a completely different thing.
The number of people who want performance stats, or at least think they need, far outnumber the amount of people who need performance. It makes absolute sense to sell these lemons, even though it is unethical.
The only part that matters. Maybe it's even illegal.
Irrelevant. If you pay premium for performance expecting to get the level of performance you've paid for then you better get what you've paid for.
Otherwise we're dealing with something very similar to consumer fraud covered with a thin paternalist veil of "corporation knows best".
That makes no sense. "Premium" strictly means "at a higher cost." You can't possibly get what you have paid for with premium products, because some of your money has been spent purchasing the premium.
Regardless of sour reviews such as this one, most people will continue to buy Apple products because Apple sales have managed to transcend the quality of their products. Apple will continue to push the boundaries of what is acceptable and people will continue to buy their products, defend them and praise them - no different from the past decade.
The fault lies squarely with consumers. Apple is just doing what they are being asked to do: flatter things, bigger numbers, higher prices.
The full value out of it includes:
1) It not being short-lived by rampant overheating (even if it's within cpu tolerances) 2) It being quiet as you expect a MBP to be
It's not supposed to render 6K 3D all day at full speed in the tropics...
"You're using it wrong" (ie, "it's not supposed for" with different wording) has been a long time excuse for many Apple failures at this point and I'm tired of it.
Hyperbole aside, their advertising shows it hooked up to a 5K external display running the Unity debugger. So it should at least be able to render 5K 3D for a typical work day.
>their maximum FP32 compute performance is 11 TFLOPS (which points to around 1340 MHz clock-rate for the Vega 64) and their peak memory bandwidth is 400 GB/s (indicating about 1600 MT/s memory speed), which is slower when compared to the Radeon RX Vega cards for desktops. The main reasons why Apple downlocks its GPUs are of course power consumption and heat dissipation https://www.anandtech.com/show/12152/apple-starts-imac-pro-s...
Sad to see that this is becoming a recurring theme from Apple's pro line.
Of course, if the component still overheats, that completely beats the purpose of the downclocking.
It’s just a pragmatic approach to a physical limitation. The right choice given the clientele— debatable. But just the same it’s not unique to computer manufacturers.
Stockfish (chess engine) will batter a CPU like that.
https://www.notebookcheck.net/Lenovo-ThinkPad-T470p-Core-i7-...
I finally ditched it a few weeks ago, finally replacing it with a Ryzen 7 desktop.
I'll admit my XPS 15 was a factory refurbished model, so maybe I just got a lemon. It got so hot the battery is now swelling and it popped out the trackpad (I think there might be a recall; or if not take it in since that's gotta be under warranty).
My current laptop is a HP Spectre. It's thicker than the XPS, but I think it's got a much better cooling solution and preforms better. I don't think I'd recommend a Dell XPS again.
> Dell charges just as much if not more than Apple
Not even close. XPS 15 with 6 core, 16GB RAM, 256GB SSD is $1500. MBP is $2400 for the same specs.
What I really like on the XPS 15 is the 97 WHR battery. I haven't seen any other notebook with this capacity.
I think it's a fixed issue, but only in 2018 models, so YMMV.
It's quiet if I have it in "silent" mode, but then I might as well be using a 350$ machine, as far as performance is concerned.
Maybe you have a slower CPU, or they fixed the thermals in your model?
The fan is running at around 10%, barely noticeable.
The fans do spin up and they're audible, but not louder than my Thinkpad or any of the Macbook Pro's we have at work. It is dead silent unless I'm actually doing intensive work.
It seems people have a lot of issues with older models though, so YMMV.
Perhaps the 9570 is improved, or as many suggest Dells QC is poor and I got lucky.
[1]: https://nickjanetakis.com/blog/transform-a-toshiba-chromeboo...
No replacing of soldering paste on my part was ever able to improve the situation, with the result of having powerful machines only on paper, that were brought to their knees even by light workloads.
My Mac becomes impossibly hot and slows down playing games like Hearthstone and Civ5, FFS.
I have long decided that my next laptop (regardless of the brand) will have integrated graphics and a medium TDP CPU, and I will just limit heavy workloads and gaming to my desktop.
You get more bang for your bucks this way, too. Desktop hardware costs a fraction of a comparable laptop.
The tradeoff is convenience, where you would have single machine that is able to fill all your needs (power and mobility), instead of two.
But, for me at least, in the end it's not worth it. Also, cloud and sync services made working in parallel on multiple machines almost painless.
Well, other sites have done benchmarks... have any of them reported it being slower than last year’s model?
In most, if not all, of his reviews he makes a point to check thermals and throttling related to heat, because that matters to the gaming segment of his audience.
As for media creation, he always tests Adobe Premiere for his cross-platform media creation tests while acknowledging that Premiere on Mac is less performant than it is on Windows, and that there's a significant number of Mac media creators who prefer Final Cut.
The better question to ask is if any of those other sites who did benchmarks did heavy workload testing on the computers for extended periods of time. Running a CPU hot for short bursts can yield different comparative results than for a task that takes 30 minutes or more.
I would expect so—benchmarks often have long-running tasks; if a task only takes a few seconds it's hard to distinguish the overhead from the task itself.
But if nothing else—they will do so going forward.
Yes, but depending on who you're talking to, "long-running" could mean a few minutes or several hours.
Throttling doesn't usually have a major impact on tasks that take a few minutes. But on heavy, long running loads, the effect of throttling starts to add up quickly, which is why the i9 MBP couldn't outperform the 2017 i7 MBP until it was stuffed in a freezer.
While the use of Premiere might raise eyebrows, I don't think it matters in Lee's testing, because he ran the same test on both the 2018 i9 MBP and a 2017 i7 MBP. Either way, it was a heavy, taxing, long running task.
I don't think anyone will dispute that the i9 has significant performance gains for people whose CPU utilization spikes are relatively short -- and that might be the majority of people who will opt to buy an i9 equipped MBP.
For content creators and other users who need to tax the processor for long periods of time, however, those users might be better off with the i7 version of the MBP.
With the liquid metal paste the CPU temps dropped 10°C, ran 200MHz faster and the MacBook Pro didn't thermal throttle. The standard thermal paste also saw a 7-10% improvement in thermals compared to Apple's paste.
I wonder if replacing the paste with a better one would help with the latest i9 MacBook Pro?
Former employees of Apple I know recommend reapplying the thermal paste immediately after getting a MacBook Pro.
I'm guessing their thinking is they want to avoid air at all costs, so requiring way excessive amounts ensures it will never be lacking, which they probably think is significantly worse than excesses.
[0] the difference in thermal conductivity between an entry-level thermal grease and a high-quality one is 2x~3x (entry level is ~2W/(m.K) while top-of-the-line might reach ~8W/(m.K)). The difference between air and a low-end thermal grease is 2 orders of magnitude (~0.025 W/(m.K) versus ~2)
https://www.youtube.com/watch?v=rAid5G30-WM
LM is conductive, so you need to be much more careful with the application, as over-application will kill the chip. It also can migrate over time, especially if you put a bit too much on, and is particularly unsuitable for laptops that are going to be moved around a lot.
Putting LM on a laptop is a gimmick, and one with fairly considerable risk to the hardware, not something that is suitable for mass production.
I recently re-applied thermal paste (an old Arctic MX-2 I had lying around) on a late 2013 MacBook Pro which has bare-die CPU and GPU. The first time I did it, I applied too much paste and it had an observable negative effect on thermal dissipation. Once I applied the correct amount, I got slightly better results than stock paste.
The effect on thermals are absolutely not insignificant. Thermal pastes are actually extremely bad at transferring heat and too much of it really has negative effects.
Yes, it is safer to put too much than too little. Using too much the worst case is just that the CPU throttles and performance is sluggish and that rarely leads to an RMA.
From https://www.ekwb.com/blog/thermal-compound-guide/
For example, the thermal conductivity of a high-grade thermal paste is 8.5 W/mK, and the heat conductivity of copper is 385 W/mK, or for aluminum 205 W/mK. You can see that the thermal compound is actually a poor heat conductor and that is exactly the reason why you only need a very thin layer of paste to fill the micro imperfections between the IHS and heatsink.
This has always been my rule of thumb whether it's metallic or non-metallic paste. I throw on a pair of vinyl gloves, squirt a dab on the heatsink (not the CPU), and spread an even layer with my finger. I go back and fill in low spots with a smaller dab, trying for an even surface. This way, it's thin enough that it doesn't squirt out the sides when it's installed, but it's a uniform spread that covers any gaps between the heatsink and the CPU surface. Also, by putting it on the heatsink instead of the CPU, I don't risk accidentally getting paste on the non-contact parts of the CPU.
I've also used the semi-solid thermal pads similar to what comes from the factory on most OEM heatsinks, but they are far less efficient than paste.
Page 4, http://www.arcticsilver.com/pdf/appmeth/int/vl/intel_app_met...
anyone ever friction soldered a heatsink to an IHS ?
Edit: selling custom IHS integrated heatsinks with LM applied and nail polished is a small and probably fairly inconsistent business to invest in buying a well binned 3790k kit, very few people would upgrade till an 8700k which showed significant gains worthy of an upgrade you would run dry for a few years especially the premium you would ask for such a service people would use these things for a long long long time i know people who are still using i5 2600k just because majority games are poorly multithread optimized
AMD could collaborate with EKWB would be great TR2 Liquid and Navi 96 Liquid. Intel have too much ego to collab
https://www.overclockers.com/heatsink-mounting-pressure-vs-p...
Modern chips seem to have sensors specifically detecting hotspots. E.g. with an AMD Polaris GPU, you can see "Hotspot Temperature" in HWiNFO.
So, I've repaired and cleaned many, many laptops and computers over the years. For bare die components (CPU/chipset) I use something like "half a grain of rice per finger nail area". This is already plenty. I sometimes checked after mounting, and there was always some minor squeeze-out. (For desktop CPUs I also use just half a grain of rice. The higher mounting force results in about a ~3-4 cm dia circle; this doesn't extend all the way to the corners of the IHS, but it doesn't have to, since nothing is dissipating heat there.)
> there is no real downside to an enthusiastic application of paste except for making a mess
That's not my experience.
With laptops, the pressure applied by the cold plate is not that much and thermal pastes have high viscosity. (The cold plate is often not that stiff, too, so it tends to flex). Applying too much paste results in a too thick layer of paste, which results in higher temps, hotter laptops, fans spinning faster and thus more noise and annoyance.
Indeed often I found the issue with manufacturer-applied paste was that it was simply too much. I think the reason they do this is that it's harder to get wrong and it makes the product just more annoying, not break it. If they did it properly, there would be some chance of getting it wrong, resulting in a broken product. For the same reason they're using thermal pads wherever they can get away with it: They're practically impossible to misapply.
Is this stuff electrically conductive?
> While much safer than electrically conductive silver and copper greases, Arctic Silver 5 should be kept away from electrical traces, pins, and leads. While it is not electrically conductive, the compound is very slightly capacitive and could potentially cause problems if it bridges two close-proximity electrical paths.
If you're using a "liquid metal" compound, you probably want to be very careful, as it is conductive.[1]
A correct answer takes context into account. In this case, "vendors that always tell to put a huge amount of thermal grease on the CPU".
Now, in real life, CPU, GPU and heatsink don't have the perfect, more-than-mirror finished surface. There are small gaps and small peaks that prevent a full contact. And those gaps are filled with air. Air is a bad heat conductor. So, to reduce this bad, bad air effect, you fill the gaps with thermal paste.
Thermal paste will be better than air pockets. Thermal paste will be worst than metal/metal (CPU/GPU/heatsink) contact. So, put as little as thermal paste as needed, just to fill the gaps.
But put too few paste, or don't cover everything, and you'll end up with more gaps: the paste will prevent those metal/metal peaks contacts and if not spread evenly it will leave bigger gaps.
If you can't know the ideal, it's better to have a little more paste (but covering everything) than to few or badly spread (leaving bigger gaps).
https://www.overclockers.com/heatsink-mounting-pressure-vs-p...
I really don't feel these machines are professional level.
Future generations of MBPs didn’t have the same problems.
The point is: shit happens with any product, the question is will you be supported well? Apple’s history here has generally been good but occasionally they’ve made major blunders (eg. the iBook motherboards about 17-18 years ago).
Their size leads to amplification of issues beyond reasonable response. The keyboard issues for example are way overblown. We have thousands of MBPs and hundreds of Touchbar MBPs at our company and there hasn’t been a widespread keyboard failures.
This is a ridiculous sentiment IMO.
I have a MacBook pro next to me that is on its fourth keyboard.
Why neither Apple nor Intel is opting for better thermal compound is amazing. Tests continue to CPU temperatures drop 10+ degrees (C) just by replacing the thermal compound. Even if they opted to get the very best thermal compound, I can't see it eat into the profit of either companies.
It's not a professional machine if it's throttling heavily under load so you can't use its resources....it boggles my mind that it's even allowed to sell something like this.
It's like selling a supercar which can't go at top speed for more than 20 seconds, then it throttles speed to 60 mph.
Half the issue is the shit TIM they use, the other half is the way they attach the IHS to the CPU. It's attached with a huge amount of sealant around the edges of the IHS and simply removing it will gain you a good 10c improvement because you're moving the IHS a few mm closer to the dye, combine it with good TIM and you're golden, replace it with some liquid metal TIM and you're off to the OC races.
I could maybe forgive them if this was an isolated incident on one generation of chips but this has been the case for about 4 years now.
Delidding, shaving the adhesive, and replacing with fresh TIM produces nearly as good a result as using LM (without the risks of LM). And using LM without shaving the adhesive only produces a few degrees of improvement.
You've said this in two threads that I've read and I truly do not know what you mean by it. Is it a bad thing that Dave doesn't scream at his viewers when making a video? Or do you mean something else entirely?
These things just don't have adequate cooling.
I think throttling in this scenario is a reasonable tradeoff, much more so than overengineering the cooling solution to handle it and thereby increasing cost, weight and bulk. Sit in the shade.
Other laptops work in the sun.
EDIT: Yes, I checked, the MacBook Air has an operating temperature of 10-35c and a storage temperature of -25-45c, which means that it's technically unusable or out of warranty if you use it in a hot country on a hot day. Like in the tropics on a beach for example.
> you cannot use your laptop unless you're indoors in air conditioned comfort
Maybe it's just me being used to colder climate, but I feel that laptop's thermal throttling is going to be a relatively small issue if trying to work in sunlight when air temperature is +35C out there.
I'd really seek some shade even when it's merely +25C...
I typically work from a hammock, in the shade, until it gets to 90-95ºF. The throttling happens even when the temperature is in the range of 80-85ºF and CPU load make it much worse. It's not uncommon for me to have to seek air conditioning if I'm running a large test suite or rebuilding docker containers, else my machine throttles to the point that I can't even use Safari.
I've never had this problem with any other laptop I've owned, including previous generations of MacBooks and MacBook Pros.
I do not expect the manufacturers to overbuild the thermal capacity such that this isn't a problem, not when we're talking about portable devices which (appreciably) prioritize compactness and lightness.
In the past I would have to shade my laptop while using it to play MP3s from the passenger seat with the convertible top down while on road trips through Nevada, to prevent if from overheating in the direct sunlight. It never crossed my mind that this was some kind of failure on IBM's part, the machine was already pushing the limits of what was possible at the time in the interests of being small and light.
As for the daylight, the sun is not the same as daylight. Devices have tolerances for heat, tropical heat in Miami is not the same as daylight in Chicago.
To be clear, I don't make a habit of working in direct sunlight. Screen glare makes that impractical, even if the temperature warrants it. I do generally work from a hammock in the shade, and I deal with thermal issues constantly.
Even in the shade, at around 85ºF, I find that I start throttling noticeably when I run large test suites, recreate complex python environments, or build docker images. I keep a close eye on my system's sensor temps through iStat Menus, and try to tailor my laptop's workload around it. For example - I use Slack on my iPhone or iPad instead of on my laptop, and keep Safari closed unless I'm actively testing something in it.
Still, it seems like my machine throttles to uselessness right around noon. That's when my battery hits 20% or so unless I've been plugged in to my external battery, and about the time the temperature hits 85-90ºF.
I'm going to try using my linux desktop when it happens next, that at least has adequate cooling.
That’s literally the entire case. There is no part of a laptop designed to never touch human skin.
I've found Intel Power Gadget [1] which looks like it might help, however, I'm not really sure what these graphs should look like when throttling vs not.
[1]: https://software.intel.com/en-us/articles/intel-power-gadget...
On Intel chips with a 1 to 4 core boost, no cores will rise above the base clock when thermally throttled, voltage may increase, and frequencies will be lower. You can run a CPU bound benchmark until temps rise and note when the performance begins to decrease.
On desktop Intel processors, this usually happens between 70-80°C.
For example, the biggest thing affecting battery life is thermal load. The biggest thing affecting the maximum sustained power of a laptop or phone is the thermal design. Anyone can slap a cheap, powerful chip in; good thermal design is much harder.
I think people are going to start realizing this more and more, especially as powerful chips get extremely cheap.
The proper fix was to try to reseat the BGA by whatever means and change the way the heatsink was attached so that pressure on the BGA/board was even. I did that on two units and they survived for years after their first RRoD, and on a third unit right from the get go, and that one never suffered of any issue.
How much longer can we keep making the chips "faster" when the tradeoff is increasingly just more thermal issues?
Can't wait for a decent A-series chip MacBook.
https://www.cnet.com/reviews/hp-envy-x2-snapdragon-835-revie...
(not saying I'm convinced an Apple ARM would actually outperform the x86 but that waving around Snapdragon machine doesn't prove your point)
It used to. The problem here is that they have not been able to ship next gen chips and so they started shipping a 6-core chip on the same process as the old 4-core, and that doesn't work so well.
Also, lol if you actually think TDP means anything anymore. Over the last 5 years it's become a marketing number more than anything else. It gives you a vague category of what power consumption will look like... at base clocks, under a non-AVX workload. Modern processors will happily blast past their TDP by 50-100% if there is thermal headroom available.
(And that's where laptops come up short - they can't actually cool a steady-state load, so these laptops will happily turbo themselves into a wall, and then once the heatsink is saturated they'll thermally throttle. Looks great in a 30-second benchmark though!)
That sounds slightly made up. As you mention yourself chips pull vastly different amounts of power, and put off vastly different amounts of heat depending on workload.
> lol if you actually think TDP means anything anymore
Yes, it does not mean anything because Intel releases CPUs that exceed their ratings because they can't get their process right, which is my whole point.
If I were Tim Cook I would fire Dan Riccio, and then resign myself and hand the company over to someone who cares about products and not supply chain optimization.
Also the whole executive Memoji thing is incredibly stupid and childish. Please take your jobs more seriously -> https://www.apple.com/leadership/
I think Steve Job's own philosophy would override that assertion: Steve Jobs On The Difference Between A Vice President And A Janitor https://www.sfgate.com/news/article/Steve-Jobs-On-The-Differ...
You mean, if you were someone who cared about product, couldn't execute on it yourself, but also had the capacity to direct Tim Cook’s actions, you would cause Tim Cook to do all that.
If you were Tim Cook, you would either not have the concern that motivates this or you wouldn't need to resign to have someone who cares about product in charge.
Instead, I'd probably begin by taking a look at at Tim Cook and Jony Ive's input and approach to product design. It shouldn't be hard.
Look for the people championing the "go thinner, lighter!" "form over function" cause and act accordingly.
It’d be easier to tell if Apple offered consumers choice themselves.
That's not an option for many. If you use a desktop machine in your office, you can't ever work from home, and you can't carry your machine into meetings to demo something.
It is up to the case-manufacturer to provide adequate fans and cooling for the chip. Otherwise, Intel's chips are forced to protect themselves and start to slow themself down.
http://hrtapps.com/blogs/20180712/
Maybe the key point is he isn't running Adobe Premiere.
The thermal issues in modern laptops certainly concern me, but they've been running hot for a long time now, and being thin doesn't actually change the surface area significantly.
Intel didn't invent some new physics to get "high" performance cores crammed into a laptop, they are just pulling the same stuff the phone market has been doing for the past few years. Put enough thermal capacitance against the chip itself to absorb a few multiples of the chips dissipation, allow it to get really hot, and then throttle it hard. Hope, people keep their benchmark runs short.
There is a difference between socket 2011/2270 based i7/i9s and their insufficiently cooled, memory and IO bandwidth starved cousins found in laptops. Even a fairly low end desktop is going to stomp the most expensive laptop you can find simply due to physics. Being able to dissipate more heat means higher performance. Its possible to move the curve up/down depending on fab processes, or micro-arch, but these days there aren't any miracle cures. More power equals more performance.
Does Premiere use the GPU as well? The thermal system of the MBP needs to handle the combined thermal load of GPU an CPU.
No, but it reduces the space for a better cooling system (pipes, fans, etc).
While this obviously shouldn’t happen, it seems like something you could easily predict might be a flaw if you knew that one had to exist.
(e.g. https://www.google.com/search?q=mac+chimney , namely https://www.flickr.com/photos/idontlikewords/275104234 - some case aesthetics ;-)
I thought that was common knowledge? There's NO laptops out there that wouldn't do that (and NO mobile devices either). This is why desktop workstations are still so much faster even though they carry same frequency chips.
Does anyone know if thermal throttling is so much of an issue for iPads?
Most laptops can at least run their base clock most of the time, his MacBook couldn't even do that (2.2 GHz vs 2.9ghz). The XPS 15 is considered a PC laptop with subpar thermals and even those issues only crop up during turbo, not baseclock (same CPU).
My 10 years old HP nc6400 doesn't do that. It requires occasional (once a year) cleaning of the fan assembly though (not that hard to access). The fan is loud, the laptop is bulky but it never ever throttles. The CPU has 34W TDP, I do not think that's on the low side.
Beware of anecdata. Given the stories I hear I would squarely put the various issues in the realm of thermal paste gone bad, busted fans, or clogged airducts, and certainly not on intrinsic bad design because I have personal counterexamples of dozens of various Apple laptops from 2008 to 2016 being perfectly fine out of the box.
Obviously though, a desktop will beat a laptop's thermal performance any day but I think this is missing the point.
I find that hard to believe. I have been gaming a lot on MacBook Pro unibodies and retinas with and without an external GPUs. Especially when the internal GPU is active, the thermal load is just too high for many games and applications.
Throttling normally starts in a very subtle way by disabling clock boost and is hard to notice. Throttling below the base clock is rare and this is where it becomes really noticeable.
I did mostly only have the 15" MBPs, but thermal throttling is a fact of life on pretty much all mobile devices. I do mobile development work and having mobile CPUs run at full frequency for more than a minute or two is pretty much impossible.
Keep in mind, there's a difference between thermal throttling and low boost clocks on multi-core workloads.
Notebookcheck reviewed the i7 version which is clocked higher than mine. Max boost is 3Ghz and in their tests it maintains 2.8Ghz running Prime95, drops to 2Ghz while running simultaneous CPU and GPU stress tests.
https://www.notebookcheck.net/Review-Lenovo-ThinkPad-X1-Carb...
Rendering real time path tracing while designing a house, visualising point clouds, video editing, sound design, you can do it on a laptop but the pro market will buy a desktop because it's 'cheaper', more reliable and upgradeable
Whether they are dishonestly marketing the machine as being faster is another.
The MBP already struggled with the old i7 and with no redesign I couldn't see how they could handle the i9 without extreme thermal throttling. Turns out they can't but ship it anyway!
I quite like Mac laptops, but I've concluded that I don't really need one. I'll just grab one of the new (rumored) Mac minis to continue MacOS development as needed.
This 2015 MacBook Pro will avoid turning the fan on at any cost. I frequently watch the CPU package heat up to about 190F while the fan is still at 0RPM.
Further, the SMC chip and/or kernel are set to use fairly slow fan speeds compared to the amount of heat being produced, so even when it does come on, the system remains hotter than necessary.
But, here's the major problem: it's also become clear that there is a thermal window where the system is hot enough to cause it to throttle itself, perhaps somewhere between 140F and 180F, but not hot enough for the SMC to turn the fan on fast enough or at all, even though it could and should.
So in addition to being really hot and not using the fan to get the heat away from your lap, it gets slow at the same time, right when it's being used. The system becomes faster soon after the fan is turned up to the max.
And to top it off, there appear to be two different throttling mechanisms, the CPU speed can adjust up or down (you can see this with the Intel Power Gadget), and Apple has some kernel system that will use CPU cycles for idling instead of scheduling threads that are waiting, which reduces overall CPU usage, and therefore also heat production.
All combined, it looks like the system is constantly bouncing between ramping up the CPU frequency to handle load, then wasting some of the additional CPU time idling to avoid turning the fan on, then turning the fan on just long enough to get the heat down before turning the fan back off again, and repeat.
I just wish I could run OSX in a VM. Linux as a primary desktop isn't quite there yet.
Also, give Linux as a primary desktop a try! What do you believe that's lacking, from your perspective?
Granted, it's not great out of the box, and some software isn't available, but if you have a couple of weekends to spare/invest with tinkering, you can have a pretty and solid desktop workstation setup.
I replaced my MBP with a Surface and I’m very happy with it, so add that to your list. WSL beats my previous setup of VirtualBox easily for Linux stuff.
They ought to rename it, and actually sell a professional laptop.
Sure - there are some professions out there that require more power than a MacBook Pro can offer. But is it really fair to say it's not "Pro" just because it can't meet the needs of 100% of all working professionals?
Apparently it cannot even meet the needs of its own spec sheet. It gets throttle back below its base clock speed according to the video. That's the real issue.
Apple claims this is "More Pro" but it turns out to be just more of the same from Apple lately. One issue after another.
Soon, the beat fades and we're left in the dells
I could watch Netflix on my bare stomach with my Air. Latest Pro is quickly too hot to handle. Of course not the only issue but the computer never bounced back from that first impression.
The dedicated gfx card doesn’t even seem to run games like Civ 5 better than my Air which I’m guessing is due to all the extra screen pixels.
Don’t try that with a first gen Air. It had a very significant problem with thermal related issues. Several times a week it’d get so hot I could make pancakes off it, the rest of the time it was in a throttled state. Was a very common problem among folks I worked with (albeit mine was a personal, theirs was a work machine) too.
When watching Netflix?
Unlike newer thin laptops, they were, and still are, viable desktop replacements. Some of them even allowed upgrading the CPU. I once inserted i7-3612QM in 13” HP ProBook 4340s of a family member (it had an i3-3110M initially) also replacing fan assembly. It runs well under sustained loads, and the performance with the i7 is very close to much newer 15W mobile CPUs such as i7-8550U found in the current 13” macbook pros.
It’s a shame we no longer have these, not Apple, not Windows, nobody.
The temperature is something that bothers me but is usable, I guess.
If this is the case, I think it's just a matter of time for Apple replacing Intel with their own custom ARM chips.
Both can drive around and transport things, but the USPS Truck is purpose build to deliver mail, stop and start frequently, and be reliable. It does a few things well. On the other hand the F150 is widely produced, can be customized in many different ways, and is general purpose vehical. Could it do the function of the USPS Truck? Sure, but it would be slower, cost more to run, and be less efficient. Could the USPS Truck haul a trailer? Maybe, but it wouldn't be very good at it. Both are a truck, but both are designed to different things and excel at different things.
That crude comparison is ARM vs x86_64. ARM is getting there, and for the specific ARM functions it's more performant than the equivalent x86_64 functions. But once you go outside of what the ARM chip has been specifically designed for, your performance takes a massive hit. Will ARM be able to take off in a Macbook? Maybe in a few years, but Apple would need to dump a massive amount of R&D into their designs before I would think they would consider it (but who knows?)
P.S. if anything, ARM is the "general purpose vehicle" in your analogy. x64 is all about high throughput and high power on a chip you buy and plug in. ARM is the one where you can get an architecture license and customize the shit out of it.
[0] https://fuse.wikichip.org/news/1316/a-look-at-caviums-new-hi...
Intel is learning (or knows, depending on your perspective, see goldmont) how to build highly efficient cores too, but like ARM haven't quite figured out how to build a super high performance one that is crazy efficient. ARM's continue to be quite efficient but not particularly performant, while intel's continue to be quite peformant but not particularly efficient.
Either way, the trend is pretty clear at this point more power dissipation=more performance.
Code density is important, but you can just double the icache. Given that icache is often smaller than dcache, it seems that's not the major bottleneck.
Also, if code density is so important, why did arm drop Thumb when they switched to 64bit?
ARM can do a small subset of what x86_64 can do. Let's say for a minute ARM doesn't have a built in H.264 decoder (they both do, but that's the one feature popped into my head).
The x86_64 chip will be able to do execute the decoding of a H.264 file that faster and more efficiently than the ARM chip because the x86_64 has purpose designed instruction to do this or make the job easier.
Another Example. x86_64 can multiply and divide while ARM can't. Both chips can do the same work, but x86_64 can multiply 20*4 in one instruction while the ARM chip has to do 20 + 20 + 20 + 20, with each one being an addition instruction. ARM might be able to add faster than x86_64, but x86_64 will still be able to multiply faster. That's extremely simple, but when you get into more complex operations ARM has to spend more time doing what x86_64 can do.
> An i7 isn't going to run iOS more slowly than whatever Apple calls their latest ARM SoC.
an i7 will run iOS slowly because iOS is build specifically for ARM. Just like how PS3 emulators take a lot of effort to run on x86_64 chips, there is an emulation overhead.
> It _is_ going to blow its lid without a big ole heatsink brick on there and a fan too, which is why smartphones use ARM cores designed for efficiency.
Yeah, because it's designed to run at a higher thermal envelope. To get ARM anywhere near the Floating Point performance of an i7 you will need to increase the power and throw some active cooling on it.
> P.S. if anything, ARM is the "general purpose vehicle" in your analogy.
I don't think so, as seen above and below.
> x64 is all about high throughput and high power on a chip you buy and plug in.
Yes, because OSs have been designed to use as much as the chip can offer it. But I can get a 5 watt Intel CPU will will run circles around a 5 watt ARM chip.
> ARM is the one where you can get an architecture license and customize the shit out of it.
Yes, so you design it to fit your specific functions. Apple's CPUs only implement what the the engineers need it to, so they can save die space by not implementing unused instructions. You are proving my example here. ARM is the Customized Purpose built USPS Truck, while x86_64 is a general purpose F150. The USPS build their truck to fit their needs just as ARM Chips are built to meet the needs of the consumers.
My XPS 13 (lovely machine, thermal issues aside) has been RMA'd 3 times because it overheated. It still gets far too hot and throttles itself so I've had to underclock it and turn off "turbo" mode. Other users have mentioned that simply repasting the CPU has caused massive improvements in temps, but I'm personally not comfortable with messing around inside a $2000 laptop.
Users shouldn't have to cripple their hardware or actually fix components to make it work properly.
So maybe if you still want to use the laptop after a few years you can open it up and give it a bit of a facelift. Maybe applying a whole bunch of thermal paste will give it a slightly new lease of life?
IIRC the "recall" involved replacing the logic boards with refurbs having the same GPUs. The CPUs on the boards seemed to be throttled after the replacement.
That 20% extra dissipation power might be enough to get it from 2.2 to 3 GHz considering the base clocks are usually where processors have the most efficient marginal frequency scaling.
IMO, designing the thermal solution + chip combination to reach base clocks at typical office climates seems reasonably smart/defensible because most places will be cooler than that most of the year even before AC.
Might just be silicon lottery, might be the GPU thermals are interfering or maybe the 2018 would run faster with 2 cores disabled.
I remember buying the 2006 Macbook with a CPU upgrade. A Core 2 Duo with 2GHz instead of 1.83. Due to heat, the CPU throttled under load and ended up with the same speed as the standard model.
But on the other hand this is good that it a came out, should be fixed in next upgrade I would assume, no wonder we got huge bump spec, people voiced their concerns loudly.
I just see really different behavior with heat, fans, and kernel behavior between XNU vs Linux, and Windows vs Linux on the same hardware.
The OS can govern how the fans spin up, but all CPUs will throttle down when they hit a predetermined temperature.
Here they are in isolation compared to the 2017 and 2016 15-inch models sorted by score descending overall:
| Model | Configuration | Score |
|---------------------------------|------------------------------------------|-------|
| MacBook Pro (15-inch Mid 2018) | Intel Core i9-8950HK @ 2.9 GHz (6 cores) | 22547 |
| MacBook Pro (15-inch Mid 2018) | Intel Core i7-8850H @ 2.6 GHz (6 cores) | 21266 |
| MacBook Pro (15-inch Mid 2018) | Intel Core i7-8750H @ 2.2 GHz (6 cores) | 21096 |
| MacBook Pro (15-inch Mid 2017) | Intel Core i7-7920HQ @ 3.1 GHz (4 cores) | 15550 |
| MacBook Pro (15-inch Mid 2017) | Intel Core i7-7820HQ @ 2.9 GHz (4 cores) | 15251 |
| MacBook Pro (15-inch Mid 2017) | Intel Core i7-7700HQ @ 2.8 GHz (4 cores) | 14374 |
| MacBook Pro (15-inch Late 2016) | Intel Core i7-6920HQ @ 2.9 GHz (4 cores) | 14134 |
| MacBook Pro (15-inch Late 2016) | Intel Core i7-6820HQ @ 2.7 GHz (4 cores) | 13706 |
| MacBook Pro (15-inch Late 2016) | Intel Core i7-6700HQ @ 2.6 GHz (4 cores) | 13008 |
Geekbench takes a few minutes to run. If you ran it back to back to back without letting the laptop cool down, you'd see different numbers.
I reached out to the video author to see if he'd be willing to run Geekbench on it in a freezer vs not.
https://twitter.com/kicksopenminds/status/101942058846774067...
1. Run GeekBench back to back (realistic: a prolonged workload... which would show the Mac as being worse)
2. Working in a freezer (completely unrealistic situation... which would make the Mac seem better).
Only one suggestion gets passed on to the video author. Science (Apple flavour).
Are you sure?? That sounds a bit...hot. Do you mean 100F?
The fan is broken, so it can reach shutdown temperature even when I have the "powersave" governer set (which I normally do)
Dave Lee ran some tests that took a half hour to complete with the CPU pretty much maxed out. I would imagine the CPU was running hot for most of that time.
All modern chips thermally throttle. Run a heavy AVX2 workload and you'll often go below the so-called "base" frequency. Run one or two threads and it can dramatically "overclock". It is the nature of all current chips.
Eh. For the overwhelming majority of users I doubt this would ever be an issue. Compile a kernel. Apply a filter. Do some convolutions. Every normal thing will be ridiculously fast. Run a server or workstation style load and it won't be so good. That has been the case with laptops for time eternal.
Only ones in systems that have not been designed correctly!
The fact that the Aero x15 he tested with the 2.2Ghz base clock 6 core i7 can all core boost to 3.1Ghz in the same workload that brings the macbook pro far below it's base clock speed proves that!
P.S, it does not matter what program it is running, throttling is a hardware flaw!
I don't think Apple (or most other PC vendors) use the best thermal compound on their CPUs.
When my early 2011 MBP was having heat problems, I replaced the compound over the CPU and GPU and got noticeable although not spectacular reductions in heat.
Of course, that's not something you can do on a "modern" Macbook Pro, since they're not intended to be user serviceable any more.
The problem with all high end mobile CPU's is that they are given a tiny heatsink and heatpipes which are simply not up to the job. They quickly become heat soaked and the fans are unable to reduce the heat, so the units throttle.
Now especially with the mac business they are reducing the scope and value of their brand, and becoming purveyors of high value trinkets. This is nothing short of a tragedy.
There is a difference between a small premium for good engineering, components choice and design, and profiteering to the point of becoming an extremely overpriced luxury product with more brand than real value.
The hardware they put inside the later models has no hope in hell of running normally with all the thermal issues.
Try lesr boosters :)
https://images.anandtech.com/reviews/mac/retinaMacBookPro/pe...
I don't understand why there are two significantly different ways to do simple things like connect to a network or configure an existing connection. The OS throws the pretty/flat metro inspired applet at you, but it has 10% of the actual functionality you're looking for, so you end up popping open the control panel anyway. The start menu ads are an insane addition. The apps in the app store are pretty much pointless and anything you actually need you can install via chocolatey.
If MS were to produce an actual "pro" version of Windows that removed all of the fluffy inconsistent stuff, ripped out all vestiges of Win8, and put PowerShell front and center, it would appeal to a lot of people that just want to get stuff done.
If you think people buy MacBook Pros because of their tech specs, you are missing the point. MacBooks are by far the best general purpose, well-rounded laptops on the market from a build quality perspective. Everything from the screen to the touchpad to sleep/hibernate (and, until recently, the keyboard) is finely tuned to the point that you can't find another laptop on the market that just feels anywhere as nice as a whole. The tight integration between the hardware and the software doesn't hurt, either.
If all you need is a powerful laptop, or if you don't care about any of those details, you are probably better off getting a more cost effective machine, but you will always sacrifice some combination of things for it. (For me personally, the biggest one is the touchpad.)
I say this as someone who doesn't own any other Apple products, but I will likely never buy a laptop other than a MacBook Pro.
I do native Windows dev for a living, but I have no urge to pick up a machine with a touch screen unless it's a tablet. Is a touch pen/surface actually useful in day to day work?
As a pen input device it's at least as good as my Intuos2 which it replaced, plus you get to draw right on the screen. (I don't like the side button though, it requires too much force to press.)
I agree completely on the touchpad. I picked up an XPS13 that I'd intended to use as a replacement. The touchpad is as good as I've had on a non-Apple product, but it's still not on par with Apple. On the otherhand I've started playing with OpenBSD with ratpoison & qutebroswser and it's pretty remarkable how far you can get without a touchpad at all.
A couple years ago I wouldn't have considered anything but an mbp, but with the recent touchbar and butterfly keyboard shenanigans I would absolutely consider something other than an mbp. I hope Apple comes to recognize that reliability is more important than super skinny unrepairable machines.
Having said that, it's not like Apple hasn't had their share of lemons. I was unfortunate enough to have an early 2011 15" Macbook Pro. Without getting into too much details, it's a known lemon that has been discussed here on HN many times. My 2011 MBP was the first laptop I've ever had that actually stopped working. FWIW, most of my other laptops were PC laptops that didn't cost nearly as much as the MBP.
When was the last time you upgraded macOS? It's been free for several years. :)
Also, to be fair to Microsoft, Windows 10 is supposed to be the last version of Windows, so you should only have to buy it once and never upgrade again.
To be pedantic, a GNU/Darwin/Mach/BSD machine.
A more accurate statement is that it's a true posix compliant Unix based system down to the metal.
I have a Mid-2010 MBP that is a large and ugly paperweight.
Due to shonky quality control, it hard-locks every 5 minutes when using the GPU. This only started when Mountain Lion was released, but good luck rolling back if you want to actually use XCode.
Given that it was such a piece of garbage, and that (at the time) workers were jumping to their deaths rather than keep building Apple products, and more recently the crap with sneaky performance reductions on iPhone, I have sworn completely off Apple products.
Turning them into Enterprise friendly machines from a software perspective was a different story.
- iMessage
- excellent integration between Contacts, Maps, iMessage, Mail, and Spotlight
- Mail.app, the best desktop mail client (IMO)
- Siri > Cortana > Whatever tarball you are expected to run autoconf inside of to get your linux box to act like a mac from the 90s
- excellent HiDPI support (neither Windows nor Linux have figured this out yet)
- plug and play external display support
- configuration-free lid switch sanity
- configuration-free trackpad sanity
- Lack of vendor adware (Windows obvs, did Ubuntu ever remove those Amazon search things?)
- painless FDE (extremely important)
- good fonts
- iOS development possible
- the most common modifier key is where my thumb can reach it and is not overloaded with the unix modifier key (is ctrl-c copy or SIGINT?)
- copy and paste work between apps using consistent keystrokes (windows figured this one out, linux not so much)
- less malware than windows
- an order of magnitude better a11y than anyone else in the history of OSes (it’s literally not even a race anymore, Apple is so far ahead here)
Finally, in not-OSX:
- build quality, even with keyboard and thermal issues, is still simply a head and shoulders higher than everyone else
The only people who could reasonably challenge Apple’s level of fit and finish from an OS perspective are Google, which is why I carry an rMBP and the 16gb Pixelbook. ChromeOS these days is a fucking marvel, up there on quality par with google search before answer cards, google assistant, and google maps. It’s a major achievement and is the first actually workable/usable Linux desktop.
I run Autodesk Inventor 2018 on a bootcamped mid-2017 Mackbook 12”, it has a built in Intel video card (650 I believe). The only time I wish I had more processing power is when doing huge FEA analysis, my 15W processor goes into a down-clocked limp mode, and computations take 4X longer than on a latest gen desktop).
If it costs $3-5k and I use it for hours and hours each day, years on end, it's peanuts as a professional tool. I'd rather pay and enjoy using it than take a discount but be frustrated each day.
My workflow runs faster on many laptops outside the walled garden. But I've gotten used to an exceptional build quality that's reliable under hard use, and OSX lets me think & work more like I want than other OSes. Also, my customer experience with Apple has been outstanding for decades.
For me, Apple has a mix of quality and aesthetics that I like better than what I've experienced from other manufacturers. It's been worth the price premium for me.
That said, I'm avoiding the butterfly keyboards until they really get them figured out.
Yes, mac touchpads are light years ahead of non mac touchpads.
- running virtual machines.
- running MemSQL which requires a minimum of 8GB to be available.
- running video editing software.
There are probably other reasons, but those are the ones off the top of my head.
I bought an Intel Hades Canyon with 32GB to run Linux on after the new MBPs didn't turn up at WWDC. Feeling somewhat relieved right now.
In normal day to day use most people will not do enough processing to be affected by thermal issues.
Keeping the form factor of a MacBook is much more important than eliminating every thermal edge case. People take for granted how much we’ve improved on old big and heavy laptops (with worse performance even).
So, making it an inch or two thicker?
https://www.youtube.com/watch?v=wgeh7ZJRhZU
edit: Bear in mind these videos are about new MacBooks so there's no way to predict their usual lifespan. And I'm not saying all of them die after 2-3 years. Apple is simply optimizing for the 90% of users here, people doing mostly web browsing and using Microsoft Office. This does not put a sustained load on the CPU/GPU so these users will be fine.
Conspiracy theory BS. MacBooks don't die after "2-3 years", and they retain much better resale value than Wintel laptops too.
That said, Apple could very well prioritise quiet over full-speed.
And yet, MacBooks tend to be extremely reliable for many years.
So that's actually pretty well optimized.
Maybe, but I've seen reliability formulas with fifth powers of temperature, namely for equivalent ageing of devices.
And 14nm fabrication processes don't exactly improve my expectations of durability ...
It's Steve Jobs' ploy to kill the laptop market. He didn't need the laptop market, he had the iPhone, he knew of the iPad when he demo'd the Macbook Air. Today the entire Mac revenue https://www.statista.com/statistics/382260/segments-share-re... is below 10%.
See, he have started the thin craze and everyone copied. But noone needs a thin laptop. There's no advantage to it whatsoever. People do need lightweight laptops and well before the Macbook Air we had shockingly lightweight ones, my Panasonic CF-Y5 was 1.5kg including a DVD ROM more than ten years ago. It had a protrusion on the lid to create a cushion which made a review to call "the exterior design of the machine's casing is reminiscent of a Sherman tank cross-bred with a 1970s sports saloon". It made it 45mm thick at the thickest point. Still, it was 1.5kg and undestructible.
So, instead of focusing on materials science and building lightweight laptops in whatever thickness a good cooling system would require everyone is falling over themselves to produce a thin laptop. Which, of course, can't be properly cooled. It's just physics. In this case, the Macbook "Pro" is 0.61" high where the Lenovo P52 (which is Pro for real) is 0.96”. This thinness craze killed off the convenient bottom dock ports from the business laptops creating these awkward side docks.
Fantastic trick, Mr Jobs, well played.
I own an Air, and it's been my go to machine in university. Yes, the weight does a lot to help me just pick it up and grab it, but the thinness does help a ton too. Thin means that I can throw it in a bag and still fit textbooks, notebooks, hoodies, and whatever else inside. I've had bags packed fully and needed every inch of space. The Air does that.
It's not a huge market, and I think only the Air models should focus on it, but it does exist. A bigger but equally light laptop wouldn't make nearly the difference that the Air does in how it fits ends up being the smallest thing in many of my bags.
He just had a great idea and Apple executed on it. The early MBA's (2010) were absolutely fantastic, quiet and reasonably powerful (read completely usable).
As an anecdote, new hires at my company are given either a somewhat heavy ThinkPad or MPB (higher end). Many have complained about the weight of the ThinkPad. None complain when they choose the MBP.
My macbook pro 2016 i7/16G doesn't even skip a beat.
Unless you're doing 3d graphics or Machine Learning or something...I think for 99.9999999% of POWER USERS a throttled i9 with 32 gigs of RAM is plenty for the year 2018.
This throttle issue in my opinion is such a minor quibble it feels like demonization of Mac.
You can focus on the very very very few negatives of Macbooks but in my opinion at the end of the day it's an incredible machine.
99.9999% of people will never go 300mph..unless you're using it in a very specialized way.
No laptop can be an equal to a real workstation unless you don't actually need a workstation to begin with.
It's not much different than my iPhone shutting down after recording 4k60 video for 2 minutes in the hot summer sun at 110F degrees (in fact, that's actually more frustrating than this Macbook "issue" but there was no outrage for that).