New MacBook Pro Features Smaller Heatsink Due to Supply Chain Issues

[panjandrum]

Once again, an issue that will not affect 99.5% of users is being treated as the freaking apocalypse.

Poor design is poor design; the number of people it directly impacts isn't a great way to measure that: "The Tesla 'self-driving' mode murders less than .1% of owners! A non-issue since it's such a small number!"

More importantly, IF it is a problem (we don't know yet), it will impact precisely the group of owners who actually *need* that performance. It's a supposedly 'pro' machine, right? So even if we are judging the design by users impacted, that's more like "impacts 100% of users who actually need a MacBook Pro."

We have a TON of the M2 MacBook Airs, and their performance is fine; but when I run extended tests that push both the CPU and GPU they are only faster than the M1 MacBook Air models for the first few minutes. After that the M2s thermal throttle so badly that they are outperformed by the supposedly slower model that they replaced. Absolutely no issue for us in the hands of K-8 students, and they aren't 'pro' labeled, but even though it doesn't impact us it's still a bad design. This is something Apple has been doing for ages, and ages, and ages, and ages BTW; undercooling their equipment. It's probably why you are seeing these reactions; we know Apple's history in this regard, and it's not exactly a glowing one. Or rather, it is a glowing one - glowing from all the excessive heat.

I'll be interested in seeing how the M2 MacBook Pros benchmark in extended tests against the most-similar outgoing M1 models. Cinebench and Heaven (or similar) running simultaneously in long throttling tests. Hotter is OK, as long as it isn't enough hotter to cause excessive throttling.

Regardless though, until we have some concrete results we won't know if this is even an issue.

 

[eicca]

Poor design is poor design; the number of people it directly impacts isn't a great way to measure that: "The Tesla 'self-driving' mode murders less than .1% of owners! A non-issue since it's such a small number!"

More importantly, IF it is a problem (we don't know yet), it will impact precisely the group of owners who actually *need* that performance. It's a supposedly 'pro' machine, right? So even if we are judging the design by users impacted, that's more like "impacts 100% of users who actually need a MacBook Pro."

We have a TON of the M2 MacBook Airs, and their performance is fine; but when I run extended tests that push both the CPU and GPU they are only faster than the M1 MacBook Air models for the first few minutes. After that the M2s thermal throttle so badly that they are outperformed by the supposedly slower model that they replaced. Absolutely no issue for us in the hands of K-8 students, and they aren't 'pro' labeled, but even though it doesn't impact us it's still a bad design. This is something Apple has been doing for ages, and ages, and ages, and ages BTW; undercooling their equipment. It's probably why you are seeing these reactions; we know Apple's history in this regard, and it's not exactly a glowing one. Or rather, it is a glowing one - glowing from all the excessive heat.

I'll be interested in seeing how the M2 MacBook Pros benchmark in extended tests against the most-similar outgoing M1 models. Cinebench and Heaven (or similar) running simultaneously in long throttling tests. Hotter is OK, as long as it isn't enough hotter to cause excessive throttling.

Regardless though, until we have some concrete results we won't know if this is even an issue.

If it’s only affecting 0.5% of the entire user base, it’s not a poor design.

 

[panjandrum]

If it’s only affecting 0.5% of the entire user base, it’s not a poor design.

Of course it is, IF that 0.5% of your user base are the intended users. If I build a valve adjustment tool for Porsche owners who need that specific tool, but it has a flaw where it only works correctly for VW engines, then it's a poor design. We don't even really know if this is going to be an issue, so it's not worth arguing about, but fanboying for Apple doesn't do anyone any good; it just lets Apple get away with poor decisions. My argument would be that IF this turns out to cause throttling, as is certainly possible, that it's a poor design on Apple's part because pro users; the target market for pro computers, requires chips running with as little throttling as possible.

 

[eicca]

Of course it is, IF that 0.5% of your user base are the intended users. If I build a valve adjustment tool for Porsche owners who need that specific tool, but it has a flaw where it works as correctly for VW engines, then it's a poor design. We don't even really know if this is going to be an issue, so it's not worth arguing about, but fanboying for Apple doesn't do anyone any good; it just lets Apple get away with poor decisions. My argument would be that IF this turns out to cause throttling, as is certainly possible, that it's a poor design on Apple's part because pro users; the target market for pro computers, requires chips running with as little throttling as possible.

A poor design is poor if it affects more like 20% of the intended users. If 99.5% of the intended users aren’t going to have an issue with it, and in this case they’re not, then it’s not a poor design.

 

[phenste]

The text in bold proves that you are not of the affected audience. Please be a dear and stop talking like those that ARE of the affected audience don't have anything to worry about here. If you bought an M2 Pro Mac and are only browsing the web, you obviously won't have a problem. The fears come from those putting this machine under load.

this is one hell of an essay for a web forum, I do have to laugh, I’m sorry. also, please be a dear and cut it with the passive-aggressive BS 🤭

you clearly haven’t seen any other posts of mine around here (your username is fitting given I’m practically a daily poster at this point)—I will also be using my machine for music production, forgive me for having it for roughly eight days. if my computer gets warmer, so be it—is everyone already this spoiled on Mx Macs?

thank you for inspiring a future test of mine, you can count on me making a thread about it: 100 tracks in Ableton all running simultaneously. if I’m proven wrong, all is forgiven. if not…it’ll be hilarious to see your presence (or lack thereof) in the thread.

peaches & cream & applesauce 💜

 

[IconDRT]

Folks, please, the endless bickering needs to cease. We are all friends here and we all want what's best for Apple and its shareholders. This entire controversy is only a problem if you can't afford a Mac Studio Ultra or Mac Pro.

 

[aeneasvb]

How does the M2 "heat issue" compare to an intel-based mac? If it's still running better than those one, then it's still a worthy upgrade, just not for those with an M1 mac.

 

[imageWIS]

Once again, an issue that will not affect 99.5% of users is being treated as the freaking apocalypse.

There such a thing as principal, a well specced M2 14" MBP is around $4,000... for that price the computer better be better than the previous model, not worse.

 

[imageWIS]

Folks, please, the endless bickering needs to cease. We are all friends here and we all want what's best for Apple and its shareholders. This entire controversy is only a problem if you can't afford a Mac Studio Ultra or Mac Pro.

I mean, I have a fully specced out Mac Studio M1 Max (Ultra was too much $$$ for my needs) which wasn't inexpensive, but for what Apple is charging for a well specced M2 Max 14" MBP (around $4,000), it damn well better not heat needlessly because Apple wanted to save 50 cents! I almost bought an M1 Max 14" MBP with the same specs as my Mac Studio, I would have been furious if it heated up significantly more than the Mac Studio...

 

[eicca]

There such a thing as principal, a well specced M2 14" MBP is around $4,000... for that price the computer better be better than the previous model, not worse.

Good news: everyone who has tested it finds it’s not worse.

 

[paulchiu]

The reason for the use of four smaller memory modules appears to be supply chain issues. The entire SoC is mounted on a substrate, so four smaller modules allow Apple to use a smaller substrate, making a saving on materials and reducing complexity as a result. Dylan Patel, Chief Analyst at SemiAnalysis, told iFixit:The M2 Pro and M2 Max deliver up to 20 percent better CPU performance and 30 percent better GPU performance than their predecessors, but since the chips continue to be based on TSMC's 5nm process, some users have noted that Apple may have made thermal tradeoffs in order to deliver improved performance.

Article Link: New MacBook Pro Features Smaller Heatsink Due to Supply Chain Issues

Who said that the use of smaller heat sink was due to supply chain issues?
A source/ a person from Apple itself, a third party vendor, care to quote here?

 

[kyykesko]

P.S. To the folks complaining about smaller heatsinks... I really don't understand you. You get a faster computer in all scenarios AND a longer battery life. Why do you care about how large the heatsink is if the computer performs extremely well?

It's a LAPtop. If you use it on your lap you may care how hot it gets. The better the heatsink the faster the heat dissipates, and maybe, just maybe your balls won't boil. Do you understand now?

 

[kyykesko]

Somehow .. somehow I doubt that you're running a workload that will generate this kind of heat while you're using it on your lap.

It's nice to know we can always ask from you how we are using our computers as you seem to know it better than we do ourselves.

 

[leman]

It's a LAPtop. If you use it on your lap you may care how hot it gets. The better the heatsink the faster the heat dissipates, and maybe, just maybe your balls won't boil. Do you understand now?

Dissipation is done by the radiators (aka. spinning fans). The job of the heatsink is merely to transfer the heat from the chip to the radiators.

Third-party reviews show that the new laptops do not hotter or louder than the old ones. So what's the problem exactly? Maybe Apple improved the performance of the heatsink (some new heatpipe design etc.), or maybe the old heatsink was massively over-engineered for the little heat these systems produce.

 

[Aenean144]

Who said that the use of smaller heat sink was due to supply chain issues?
A source/ a person from Apple itself, a third party vendor, care to quote here?

> Who said that the use of smaller heat sink was due to supply chain issues?

It is Macrumors' interpretation of iFixit's article. It's blatantly false as neither iFixit nor Macrumors understands what a "heatsink" is, and everything in iFixit's article is a guess regarding reasons for why Apple does things.

What they are referring to as a "heatsink" is the hot side interface plate of the heat pipe going from the SoC to the radiator fins. If someone says heatsink, you should be thinking radiator fins as your first thought, or perhaps ambient atmosphere.

So the question is really, "why is this heat pipe interface plate different?" Really, nobody knows. What I see there is absolutely fine and should work identically to the M1 Pro version and any other interface plates or blocks from the gazillion coolers on the market. The plate has to cover the largest heat source, the M1 Pro chip, and it does. It's absolutely fine.

The interface block should not cover the RAM because they run a lot cooler than the logic chips and if you bridge the RAM with the heat pipe interface block, it could heat up the RAM, not cool it. That would be a problem.

> Who said the heat sink is smaller?

Follow the links. Shahram Mokhtari of iFixit said this: "The most interesting of which is the apparently reduced size of the heatsink on the M2 Pro SoC. This prompted a bit of investigation since we would have expected the M2 Pro to have a similar if not larger heatsink compared to the M1."

So, you can blame it directly on him. He's an all-in fix it yourself person, or is employed to be. Does a nice job taking things apart and putting it together, but he is Youtuber-like, who doesn't really know anything about how things work, and loves to throw out controversy bombs for the engagement.

He asks Dylan Patel why Apple would go from 2 big RAM packages in the M1 Pro versus 4 smaller RAM packages in the M2 Pro. Patel is quoted as saying "ABF substrates were in very short supply when Apple made the design choice. By using four smaller modules rather than two larger ones, they can decrease routing complexity within the substrate from the memory to the SoC, leading to fewer layers on the substrate. This allows them to stretch the limited substrate supply further."

I don't know if I buy either of these reason from Patel regarding the RAM configuration, and he's just guessing on what Apple's supply chain could be. [The bigger RAM packages in the M1 Pro and M2 Max actually contain those 2 RAM packages seen in the M2 Pro. The maximum capacity mass produced LPDDR5 is 12 Gbyte, which fits in these smaller looking RAM modules in the M2 Pro of M1 and M2 SoC. The M2 Max has 4 of these larger packages, inside each are 2 LPDDR5 RAM packages. 4 x 2 x 12 = 96. That's how the M2 Max MBP models support 96 GB RAM.]

Anyways, I digress. You know what Patel didn't say or is quoted about? The size of the "heatsink". Mokhtari probably threw out the "heatsink" word because he's an idiot about it, never thinking what heatsink really means, and really just tried to explain the 4 versus 2 RAM modules. It's basically craziness to include RAM as part of the heat transfer chain of the CPU and GPU.

 

[leman]

He asks Dylan Patel why Apple would go from 2 big RAM packages in the M1 Pro versus 4 smaller RAM packages in the M2 Pro. Patel is quoted as saying "ABF substrates were in very short supply when Apple made the design choice. By using four smaller modules rather than two larger ones, they can decrease routing complexity within the substrate from the memory to the SoC, leading to fewer layers on the substrate. This allows them to stretch the limited substrate supply further."

I don't know if I buy either of these reason from Patel regarding the RAM configuration, and he's just guessing on what Apple's supply chain could be. [The bigger RAM packages in the M1 Pro and M2 Max actually contain those 2 RAM packages seen in the M2 Pro. The maximum capacity mass produced LPDDR5 is 12 Gbyte, which fits in these smaller looking RAM modules in the M2 Pro of M1 and M2 SoC. The M2 Max has 4 of these larger packages, inside each are 2 LPDDR5 RAM packages. 4 x 2 x 12 = 96. That's how the M2 Max MBP models support 96 GB RAM.]

The RAM packages contain multiple RAM chips, possibly stacked (or in some other configuration). But regardless of whether it's 4 or 2 packages, it's still 256-bit RAM bus and thus I don't really understand how routing complexity would be different. Maybe it's more space-efficient or maybe it was cheaper to make, or maybe Patel is right and it's a way to use materials more efficiently, who knows. The RAM packaging is custom anyway.

Wha't really funny if any other vendor changes the heatpipe configuration everyone goes "uuh, company X is using VAPOR CHAMBER cooling", but if Apple does it it's "damn Apple nerfing my cooling system". Go figure.

 

[paulchiu]

I just ordered a 16 with M2 Max and 96GB, should be toasty if iFixit folks are right. This will be my third Mac with Apple M Series chips.

 

[steve123]

I just ordered a 16 with M2 Max and 96GB, should be toasty if iFixit folks are right. This will be my third Mac with Apple M Series chips.

I am sure you won't be disappointed ... it is really fast and I have no burns.

 

[paulchiu]

I am sure you won't be disappointed ... it is really fast and I have no burns.

I had happy with the 2021 MacBook Pro M1 Max with 64Gb. This new one will get even closer to functionality of our 2019 Mac Pro.

 

[sxlx]

I am seeking information regarding feasible modifications for the heatsink, which could take the form of either a replacement or additional components.

 

[leman]

I am seeking information regarding feasible modifications for the heatsink, which could take the form of either a replacement or additional components.

To which end? What do you want to achieve?

 

[sxlx]

To which end? What do you want to achieve?

Well, it's clear that the purpose of a heatsink is to dissipate heat, wouldn't you agree?

 

[leman]

Well, it's clear that the purpose of a heatsink is to dissipate heat, wouldn't you agree?

No, the purpose of the heatsink is to transfer heat to the radiators. The reason why I am asking is because you obviously have something specific in mind, and I have difficulty guessing what it is. I mean, the current heatsink is more than good enough to deal with any amount of heat these chips are able to produce, so I just don't see why you would want to upgrade it.

 

[sxlx]

No, the purpose of the heatsink is to transfer heat to the radiators. The reason why I am asking is because you obviously have something specific in mind, and I have difficulty guessing what it is. I mean, the current heatsink is more than good enough to deal with any amount of heat these chips are able to produce, so I just don't see why you would want to upgrade it.

While the transfer of heat to the radiators is an important aspect of the heatsink's purpose, its primary function is to dissipate heat. An upgraded heatsink will not only effectively transfer heat to the radiators, but it will also ensure optimal dissipation, resulting in improved performance and longevity of the system. I believe that this proactive approach is well worth considering.

Recent findings regarding the new M2 Max chip have shown that it runs relatively hotter than the previous generation M1 Max Chip, likely due to increased power output. And to add insult to injury, there has also been a downgrade in the heatsink from the prior generation. In light of this information, upgrading the heatsink becomes even more crucial in ensuring maximum performance and longevity of the system. An upgraded heatsink will not only provide improved dissipation, but also ensure that the chip operates within safe temperature ranges, preventing any potential damage.

 

[steve123]

While the transfer of heat to the radiators is an important aspect of the heatsink's purpose, its primary function is to dissipate heat. An upgraded heatsink will not only effectively transfer heat to the radiators, but it will also ensure optimal dissipation, resulting in improved performance and longevity of the system. I believe that this proactive approach is well worth considering.

You may want to go back and review some of the earlier posts here. You are mistaking the metal slab as being the "heatsink". Rather, the entire apparatus is the "heatsink". The slab is not intended to dissipate the heat. In fact, if it did it would likely have an adverse affect on the adjacent DRAM. Rather, the slab acts as part of the heat pipe that conducts the heat to the radiators. The radiators dissipate the heat.