14" M1 Max - 100C on Cinebench R23, within seconds. YT reviewers get 51C.

[leman]

As you can see I'm replying to a post that mentions failures/problems at specific temps.
If you don't find performance drop as a problem then I guess you never ever used the computer hard enough to know what is it that's I am talking about. Time = money. That's for reading comprehension.

Intel Tiger Lake 11980HK is rated to run at 2.6ghz base frequency and 5ghz turbo. That’s almost 2x delta. The performance drop is not a problem, it’s part of Intel CPUs very design. Sure, you can run them with more potent cooling and you will get more performance, but then you are running them outside of the nominal spec. I do not agree that running a chip outside the nominal spec should be normalized.

Apple CPUs do not have such a wide range of operating frequency because of their different design, and the difference between their peak performance and sustained performance is much smaller.

As for new silicon performance throttling I'll wait till I can get my hands on apple's diagnostic tool again but till that happens I'll apply same logic and behavior I do to keep all my systems running healthy. We're talking about making money here so 18 degrees C temp drop at WORST will only make the computer feel cooler to the touch and at best save it's longevity while providing 100% of power whenever I need it. From my perspective it is a win win no matter how you look at it.

What Apple diagnostic tool? powermetrics is part of the base OS install, you don’t have to „get your hands on it“.

As to your experiments with fan control - it’s your machine, you can do whatever you want to it, it’s entirely your business. I’m just pointing out that they don’t do anything useful. I am fairly certain you won’t see any improved sustained performance because the fan would go up thievery or the other. The only effect you are achieving is that your fans ramp up earlier than they have to.

 

[Grohowiak]

Intel Tiger Lake 11980HK is rated to run at 2.6ghz base frequency and 5ghz turbo. That’s almost 2x delta. The performance drop is not a problem, it’s part of Intel CPUs very design. Sure, you can run them with more potent cooling and you will get more performance, but then you are running them outside of the nominal spec. I do not agree that running a chip outside the nominal spec should be normalized.

Apple CPUs do not have such a wide range of operating frequency because of their different design, and the difference between their peak performance and sustained performance is much smaller.

It's like you just can't see it lol.
You never ran anywhere close to a limit I guess but being that stubborn is something else lol.

What Apple diagnostic tool? powermetrics is part of the base OS install, you don’t have to „get your hands on it“.

As to your experiments with fan control - it’s your machine, you can do whatever you want to it, it’s entirely your business. I’m just pointing out that they don’t do anything useful. I am fairly certain you won’t see any improved sustained performance because the fan would go up thievery or the other. The only effect you are achieving is that your fans ramp up earlier than they have to.

powermetrics... ?


Edit. Written in a nice way for anybody to understand the concept of lower temperature = longer and better performance.
I mean why they don't just let everything throttle huh?

Data Center Cooling Technologies and Best Practices | Digital Realty

Data center cooling is critical to maintaining the integrity of your hardware. In this article we go over some common technologies and cooling best practices.

www.digitalrealty.com

 

[leman]

It's like you just can't see it lol.
You never ran anywhere close to a limit I guess but being that stubborn is something else lol.

I simply have realistic expectations and some actual knowledge, having managed hundreds of professional machines over my career.

Edit. Written in a nice way for anybody to understand the concept of lower temperature = longer and better performance.
I mean why they don't just let everything throttle huh?

Data Center Cooling Technologies and Best Practices | Digital Realty

Data center cooling is critical to maintaining the integrity of your hardware. In this article we go over some common technologies and cooling best practices.

www.digitalrealty.com

You are running a data center? I though we are talking about laptops?

 

[IG88]

Please stop spreading FUD. There is no empirical data to corroborate the myth of "high internal temperature is bad for internal components". As you say, Apple always let's their hardware run hot — because it's the efficient thing to do. If it would damage hardware in any way they wouldn't have earned reputation of being one of the most reliable laptop brands in the industry.

Where is the empirical data proving that running a laptop at high internal temperatures doesn't shorten the lifespan of the components?

 

[Grohowiak]

I simply have realistic expectations and some actual knowledge, having managed hundreds of professional machines over my career.

What were those machines if may I ask?
And when you say realistic expectations you mean throttling to you is ok is that correct?
You don't care that actual job can drag due to unpredictable performance cuts?
Long (and by long I mean days) render ques with fixed deadlines which have to be properly calculated in order to deliver the job?
By what you are saying base model macbook air is your target device.

You are running a data center? I though we are talking about laptops?

In relative terms it doesn't matter if you are managing and array of 3.5GHz servers or 1.2Ghz notebook.
They both have a limit that can be managed by appropriate cooling solution. Dude this is 101 of common sense. I'm just trying all I can to open your perception to the basic idea of correlation between proper cooling and peak uninterrupted performance. They are married to each other. Literally.

Also I'm pretty sure you started by saying:

"...Please stop spreading FUD. There is no empirical data to corroborate the myth of "high internal temperature is bad for internal components..."

It's simply impossible to ignore such statement.

 

[leman]

Where is the empirical data proving that running a laptop at high internal temperatures doesn't shorten the lifespan of the components?

I am not aware of any peer-reviewed publications targeting laptops specifically.

However, there is at technical report about embedded processors, which should in principle apply to all kind of complex semiconductor: https://www.ti.com/lit/an/sprabx4b/sprabx4b.pdf?ts=1635453161041 (section 4)

Beyond that, there is a lot of indirect evidence:

1. CPU manufacturers (Intel, AMD, Apple) consider running CPUs at 100C part of regular operation and offer full warranty for doing that (they would hardly do it if running at 100C had serious concerns for the hardware reliability)

2. Apple laptop system have been letting CPUs to run at 100C for many years, and their laptops are used by many professionals who push the systems hard, and yet Apple laptops are still considered some of the most reliable on the market. That would hardly be the case if running laptops at high internal temperatures would have serious concerns for reliability

3. Even the new much-more power-efficient Apple Silicon systems will bounce the CPU off the 100C mark. It would have been much easier for Apple to avoid that temperature limit compared to the much more power hungry Intel CPUs. But they did not. Which means that utilizing the full temperature range up to 100C has operational benefits (most likely more efficient cooling). Apple would hardly do it if reaching 100C would have serious concerns for reliability


To sum it up, given that hardware manufacturers tell say that running components at 100C is safe and supported under warranty, given that some laptops (e.g. Apple) have been designing their laptops to run with high internal temperatures under load without any apparent reliability issues, and given that there is evidence that running embedded processors at 105C or below has no noticeable impact on their longevity (over reasonable time spans), I would conclude that if one want to claim that "heat kills" (as in, 100C is bad for hardware), the burden of proof is on them and not on people who deny this claim. After all, the claim is significant in so far it contradicts practices established in the industry.

 

[Grohowiak]

I am not aware of any peer-reviewed publications targeting laptops specifically.

However, there is at technical report about embedded processors, which should in principle apply to all kind of complex semiconductor: https://www.ti.com/lit/an/sprabx4b/sprabx4b.pdf?ts=1635453161041 (section 4)

Beyond that, there is a lot of indirect evidence:

1. CPU manufacturers (Intel, AMD, Apple) consider running CPUs at 100C part of regular operation and offer full warranty for doing that (they would hardly do it if running at 100C had serious concerns for the hardware reliability)

2. Apple laptop system have been letting CPUs to run at 100C for many years, and their laptops are used by many professionals who push the systems hard, and yet Apple laptops are still considered some of the most reliable on the market. That would hardly be the case if running laptops at high internal temperatures would have serious concerns for reliability

3. Even the new much-more power-efficient Apple Silicon systems will bounce the CPU off the 100C mark. It would have been much easier for Apple to avoid that temperature limit compared to the much more power hungry Intel CPUs. But they did not. Which means that utilizing the full temperature range up to 100C has operational benefits (most likely more efficient cooling). Apple would hardly do it if reaching 100C would have serious concerns for reliability


To sum it up, given that hardware manufacturers tell say that running components at 100C is safe and supported under warranty, given that some laptops (e.g. Apple) have been designing their laptops to run with high internal temperatures under load without any apparent reliability issues, and given that there is evidence that running embedded processors at 105C or below has no noticeable impact on their longevity (over reasonable time spans), I would conclude that if one want to claim that "heat kills" (as in, 100C is bad for hardware), the burden of proof is on them and not on people who deny this claim. After all, the claim is significant in so far it contradicts practices established in the industry.

Now that we know about chips (for 1 year though as that's how warranty is in US unless you pay extra $300-400 for applecare) we need to know about solders, transistors, capacitors, cables, coatings, screen, mechanical parts, including those in components like WiFi, bluetooth, drives and etc.
After all this is what makes a computer a computer.

Can you provide data that supports the same statement as the one you provided from TI? I mean I'm not gonna go into reliability of industrial grade components vs consumer products (look it up as the difference can be staggering) but you specifically said this applies to all components so I'll go along even though it is a false statement as they are tested differently and usually have very different thermal capabilities.

"...Please stop spreading FUD. There is no empirical data to corroborate the myth of "high internal temperature is bad for internal components..."

 

[leman]

What were those machines if may I ask?

Mostly Apple laptops, used for demanding CPU work

And when you say realistic expectations you mean throttling to you is ok is that correct?

"Realistic expectation" is when I buy a laptop CPU rated for 2.4Ghz sustained performance is that it probably won't be able to run at 4Ghz sustained performance. Apple usually does fairly well here, performing 10-20% above the spec on average.

You don't care that actual job can drag due to unpredictable performance cuts?

That is a weird question. Nobody likes unpredictable performance cuts. But you are moving goalposts. No "unpredictable performance cuts" on Apple Silicon laptops using the default performance profile have been reported, and it is not clear at all that your fan curve adjustments will offer any kind of improvements.

In other words, I fail to see any relevance of this remark to what we are discussing (which is whether more aggressive fan settings will give you more performance on M1 Pro/Max CPUs)

Also I'm pretty sure you started by saying:


It's simply impossible to ignore such statement.

Well, it's good that you don't ignore this statement. You might as well treat it as an opportunity to learn something.

Can you provide data that supports the same statement as the one you provided from TI? I mean I'm not gonna go into reliability of industrial grade components vs consumer products (look it up as the difference can be staggering) but you specifically said this applies to all components so I'll go along even though it is a false statement as they are tested differently and usually have very different thermal capabilities.

No, I cannot. Can you provide data that supports your statement (that running laptop components with CPU close to 100C results in significant reliability issues over reasonable timespans)? As I wrote, the burned on proof is on you, since there are vendors that run their laptop components at high temperatures while at the same time being industry reliability leaders. There is some sort of contradiction here, don't you think? If "heat kills", who don't Apple laptops fail in bulk?

P.S. The temperature of my Intel CPU package right now is 85C. What am I doing? Writing on Macrumors. How is my laptop still alive after two years of writing on macrumors?

 

[Toutou]

Guys, what the heck are you even arguing about?

Prolonged operation at high temperatures near Tjunction does affect the longevity of both old and modern CPUs and other components around them. That's a fact. No need to argue about basic physics.

Whether the effect is significant and even measurable in the context of the typical lifespan of a MacBook (ten years maybe?), we don't know, because we don't have enough data. And the empirical data we have (i.e. millions of MacBooks still chugging along after these years or failing for different reasons, and the fact that most laptop manufacturers and even CPU manufacturers design their products devices to run at temperatures near Tjunction for prolonged periods) seem to show that no, it's not that important. We've been running laptops at 90-100°C for decades, and Apple more than anyone.

 

[leman]

Guys, what the heck are you even arguing about?

Prolonged operation at high temperatures near Tjunction does affect the longevity of both old and modern CPUs and other components around them. That's a fact. No need to argue about basic physics.

Whether the effect is significant and even measurable in the context of the typical lifespan of a MacBook (ten years maybe?), we don't know, because we don't have enough data. And the empirical data we have (i.e. millions of MacBooks still chugging along after these years or failing for different reasons, and the fact that most laptop manufacturers and even CPU manufacturers design their products devices to run at temperatures near Tjunction for prolonged periods) seem to show that no, it's not that important. We've been running laptops at 90-100°C for decades, and Apple more than anyone.

Exactly this. Electromigration is obviously a real thing. But it's simply not a concern for personal computers.

 

[Grohowiak]

Mostly Apple laptops, used for demanding CPU work

Can you provide some examples?

"Realistic expectation" is when I buy a laptop CPU rated for 2.4Ghz sustained performance is that it probably won't be able to run at 4Ghz sustained performance. Apple usually does fairly well here, performing 10-20% above the spec on average.

Can you show me the data to support your statement?
Plenty of throttling info that completely contradicts your statement in the topic I posted here.
My case no different. Performance drops while compiling xcode projects that shouldn't even be a issue. Confirmed by apple tech team and told works as intended. How's that for sustained performance?

That is a weird question. Nobody likes unpredictable performance cuts. But you are moving goalposts. No "unpredictable performance cuts" on Apple Silicon laptops using the default performance profile have been reported, and it is not clear at all that your fan curve adjustments will offer any kind of improvements.

In other words, I fail to see any relevance of this remark to what we are discussing (which is whether more aggressive fan settings will give you more performance on M1 Pro/Max CPUs)

Not reported or you just don't want to see what is in front of your eyes?

Well, it's good that you don't ignore this statement. You might as well treat it as an opportunity to learn something.

You should practice what you preach.

No, I cannot. Can you provide data that supports your statement (that running laptop components with CPU close to 100C results in significant reliability issues over reasonable timespans)? As I wrote, the burned on proof is on you, since there are vendors that run their laptop components at high temperatures while at the same time being industry reliability leaders. There is some sort of contradiction here, don't you think? If "heat kills", who don't Apple laptops fail in bulk?

P.S. The temperature of my Intel CPU package right now is 85C. What am I doing? Writing on Macrumors. How is my laptop still alive after two years of writing on macrumors?

CPU throttling is caused by heat = heat KILLS performance of the chip for the time that it needs to cool itself down.
Pretty much every chip (iPhones included) throttles. it's a fact and heat is the main reason.
M1 Air gets roughly 15% performance cut in prolonged, high usage scenarios. I'm willing to bet that M1 Pro/Max will be same just the value might be different.

Just cause something is operating it doesn't mean that it works at peak performance. lmao how hard is this to comprehend.

Also all that you are saying is ONLY valid in the timeframe of warranty which in US case is 1 year. After than no manufacturer is responsible for the damage or performance degradation caused by factors like heat. 12 months. That's it.

**EDIT just to add here if this is not the case and the new apple chips are gold then I will be super happy but I'll still use the fan to lower the temps. I prefer to sit in the 80s than 100s with hours upon hours of 100% usage that this machine will get.

 

[dgdosen]

Ambient temperature is 73C. I'm from the desktop world, getting 53c multicore is not easy, we should have known lol.

Ambient temperature is the temperature of your current environment. 73C = 163F! That'd be hotter than Death Valley. If that is the case, I'd suggest you bring your macbook into a room with a comfortable temperature of 72F/22C.

Also, you shouldn't expect your device to ever be cooler than the ambient temperature.

 

[leman]

Can you provide some examples?

Stats (mostly lengthy bayesian simulations) and various kind of data processing.

Can you show me the data to support your statement?

Sure, here is my 2019 16" MBP after roughly 10 minutes of Cinebench R23 multicore

This is default settings, no cooling mods, no messing with fans. Ambient temperature ~ 22C.

As you can see, average frequency is at or above 2.9ghz with the package power around 50W. The CPU (i9-9980HK) is rated for 2.4ghz under sustained operation, so the laptop operates 20% above the spec

My case no different. Performance drops while compiling xcode projects that shouldn't even be a issue. Confirmed by apple tech team and told works as intended. How's that for sustained performance?

Which machine? What is the package power? What is are the clocks?

Not reported or you just don't want to see what is in front of your eyes?

I though we are talking about M1 Pro/Max? Why do you bring a MacBook Air video? Of course M1 is not running at full capability there — it's a 7-10W passively cooled chassis running an SoC capable of operating at ~25W.

CPU throttling is caused by heat = heat KILLS performance of the chip for the time that it needs to cool itself down.

You really need to decide what your story is. Does heat kill laptops? Does it kill performance? Both? Neither? Because I am kind of confused here.

As to the rest, no, it's not heat that kills performance. It's the lack of adequate cooling capacity (relative to the chips maximal power output). If your concern is lack of cooling capacity, your fan adjustments achieve absolutely nothing, because you are not changing the maximal cooling capacity, just the response curve.

Pretty much every chip (iPhones included) throttles. it's a fact and heat is the main reason.

Of course every chip throttles. Every chip that uses dynamic overclocking to adjust to the workload throttles. That is part of its design.

M1 Air gets roughly 15% performance cut in prolonged, high usage scenarios. I'm willing to bet that M1 Pro/Max will be same just the value might be different.

M1 Air is a passive cooled chassis that can dissipate less than 10W... M1 Pro/Max are actively cooled laptops with state of the art cooling systems. We have sustained benchmarks using SPEC that take hours to run and show that these laptops can match high-end desktop CPUs in some workloads. Where does your conviction comes from?

Just cause something is operating it doesn't mean that it works at peak performance. lmao how hard is this to comprehend.

Of course not. Because operating at peak performance all the time is not part of sustained operation. Because otherwise every laptop would need to have a 150W cooling solution. How hard is that to comprehend?

Also all that you are saying is ONLY valid in the timeframe of warranty which in US case is 1 year. After than no manufacturer is responsible for the damage or performance degradation caused by factors like heat. 12 months. That's it.

Luckily for me I don't live in the US. Warranty on my computers where I buy them is three years.

 

[IG88]

I am not aware of any peer-reviewed publications targeting laptops specifically.

However, there is at technical report about embedded processors, which should in principle apply to all kind of complex semiconductor: https://www.ti.com/lit/an/sprabx4b/sprabx4b.pdf?ts=1635453161041 (section 4)

Beyond that, there is a lot of indirect evidence:

1. CPU manufacturers (Intel, AMD, Apple) consider running CPUs at 100C part of regular operation and offer full warranty for doing that (they would hardly do it if running at 100C had serious concerns for the hardware reliability)

2. Apple laptop system have been letting CPUs to run at 100C for many years, and their laptops are used by many professionals who push the systems hard, and yet Apple laptops are still considered some of the most reliable on the market. That would hardly be the case if running laptops at high internal temperatures would have serious concerns for reliability

3. Even the new much-more power-efficient Apple Silicon systems will bounce the CPU off the 100C mark. It would have been much easier for Apple to avoid that temperature limit compared to the much more power hungry Intel CPUs. But they did not. Which means that utilizing the full temperature range up to 100C has operational benefits (most likely more efficient cooling). Apple would hardly do it if reaching 100C would have serious concerns for reliability


To sum it up, given that hardware manufacturers tell say that running components at 100C is safe and supported under warranty, given that some laptops (e.g. Apple) have been designing their laptops to run with high internal temperatures under load without any apparent reliability issues, and given that there is evidence that running embedded processors at 105C or below has no noticeable impact on their longevity (over reasonable time spans), I would conclude that if one want to claim that "heat kills" (as in, 100C is bad for hardware), the burden of proof is on them and not on people who deny this claim. After all, the claim is significant in so far it contradicts practices established in the industry.

Did you happen to notice this part in the link you provided?

It shows that if the processor runs at 90°C effective temperature instead of the 105°C, x2 increase is useful lifetime can be projected.

Or this part

An often quoted rule of thumb in electronics reliability for capacitors is that every 10°C increase, the lifetime approximately halves. For semiconductors, it is a similar change but there is slippage at higher temperatures.

 

[mr_roboto]

Did you happen to notice this part in the link you provided?

Or this part

You need to notice the parts where they mention their study is based on continuous operation at temperature, and 10 year lifespan at 105C is a figure they think is reasonable.

Do you use your laptop computers at maximum load 24/7/365 for 10 years straight? No? Duty cycle matters.

Outside of that PDF, you need to notice that Apple doesn't even let temps reach 100C continuous. It may hit that number briefly, but the control loops seem tuned to reach equilibrium somewhere in the mid-90s.

Maybe Apple's engineers aren't quite as grossly incompetent as random armchair-expert forum posters like to believe.

 

[crazy dave]

You need to notice the parts where they mention their study is based on continuous operation at temperature, and 10 year lifespan at 105C is a figure they think is reasonable.

Do you use your laptop computers at maximum load 24/7/365 for 10 years straight? No? Duty cycle matters.

Outside of that PDF, you need to notice that Apple doesn't even let temps reach 100C continuous. It may hit that number briefly, but the control loops seem tuned to reach equilibrium somewhere in the mid-90s.

Maybe Apple's engineers aren't quite as grossly incompetent as random armchair-expert forum posters like to believe.

Another factor to consider is that the temperature sensors on the M1 are more like hotspot sensors than chip sensors - and can be expected to be hotter by about 10-15 degrees as a result.

For those who want to know more, Linus (who can can hardly be called a fan of Apple’s typical thermal solutions) explains here:

 

[Grohowiak]

You need to notice the parts where they mention their study is based on continuous operation at temperature, and 10 year lifespan at 105C is a figure they think is reasonable.

Do you use your laptop computers at maximum load 24/7/365 for 10 years straight? No? Duty cycle matters.

Outside of that PDF, you need to notice that Apple doesn't even let temps reach 100C continuous. It may hit that number briefly, but the control loops seem tuned to reach equilibrium somewhere in the mid-90s.

Maybe Apple's engineers aren't quite as grossly incompetent as random armchair-expert forum posters like to believe.

Cool so now please show us data from Intel, Apple and AMD on consumer chips.
Why we have only 12 month warranty if all that is inside can EASY work 10 years at 100%?

 

[crazy dave]

Cool so now please show us data from Intel, Apple and AMD on consumer chips.
Why we have only 12 month warranty if all that is inside can EASY work 10 years at 100%?

There are a hundred things that can go wrong with a computer that don’t involve heat destroying the internals en masse. Companies are cheap and US lets them be while other countries mandate longer warranties for silicon/computers. Plus the temperatures you are measuring are may not be what you think they are …

Edit: as the video says, if you want to run your fans at max speed for that extra sustained performance (heck the M1 Max comes with a built-in option to do that!) and peace of mind then okay, but you’re probably not increasing the longevity of your computer. In fact I’d go so far as to add to that: you are probably hurting it as the most likely thing to break will be mechanical, eg the fans you’re constantly running harder, than electrical.

 

[ADGrant]

Also MaxTech was using 16 Pro (not max). It's not logical to think 53C on 16" and 97C on 14". I do have a max though, still, that's an incredibly large difference lol.

You have twice as many GPU cores in a much smaller chassis with a weaker cooling system. You are obviously going to see much higher temps when running GPU intensive benchmarks. Apple didn't even offer Intel's higher end CPUs and discrete GPUs in the 13" MBP.

 

[Grohowiak]

There are a hundred things that can go wrong with a computer that don’t involve heat destroying the internals en masse. Companies are cheap and US lets them be. Plus the temperatures you are measuring are probably not what you think they are ..

Obviously but in this subject I'm being told hi temps in the range of 100C for extended periods of time are perfectly fine so I would love to see some solid proof from a company that you lads are defending and not industrial grade chip vendor.

+/- 2-4 hours per day at 100% with temps in 100s. That's my approximate workflow for this unit and you want me to believe that my chip will deliver same output few months/years down the road even if I let it (and everything around it) cook for hours at a time?

 

[crazy dave]

Obviously but in this subject I'm being told hi temps in the range of 100C for extended periods of time are perfectly fine so I would love to see some solid proof from a company that you lads are defending and not industrial grade chip vendor. +/- 2-4 hours per day at 100% with temps in 100s. That's my approximate workflow for this unit and you want me to believe that my chip will deliver same output few months/years down the road even if I let it (and everything around it) cook for hours at a time?

One thing is that it’s probably not hitting 100 degrees across the chip, these are hotspots that are being measured rather than the average chip temperature which is what people are often comparing it against when they say “my x processor isn’t getting that hot!”. I personally don’t have numbers for why these temps are fine over the expected lifespan of a computer. I can only report what I’ve read from a variety of sources that high temperatures (at least up to here) are unlikely to cause longevity issues. This is backed up by that report and posts from other users on this forum. Could it be a common knowledge thing that turns out to be wrong? Sure I guess, this is an area where I don’t have a lot expertise so I rely on others.

However if we accept the received wisdom above as fact, as I said in my edit, while you are certainly free to run your fans harder and longer to give yourself peace of mind, that seems to me to be more likely to cause you problems when the fan mechanism fails.

 

[Grohowiak]

However if we accept the received wisdom above as fact, as I said in my edit, while you are certainly free to run your fans harder and longer to give yourself peace of mind, that seems to me to be more likely to cause you problems when the fan mechanism fails.

This I can't argue. Fans will definitely be under more stress but never had one fail me (YET lol) so I'm ok with taking that risk.
I did modify the settings to stay inactive till 70 Celsius to imitate apples mapping (at least for now).
Small increases till 85 and then higher rpm kicks in. After todays render 81C average makes me happy. Without it I was constant 98-100.

 

[crazy dave]

This I can't argue. Fans will definitely be under more stress but never had one fail me (YET lol) so I'm ok with taking that risk.
I did modify the settings to stay inactive till 70 Celsius to imitate apples mapping (at least for now).
Small increases till 85 and then higher rpm kicks in. After todays render 81C average makes me happy.

I’ve had fans fail me … though never on a Mac (yet) and to be fair, they were old. I say do whatever makes you feel comfortable and brings you peace of mind.

 

[Grohowiak]

Another factor to consider is that the temperature sensors on the M1 are more like hotspot sensors than chip sensors - and can be expected to be hotter by about 10-15 degrees as a result.

For those who want to know more, Linus (who can can hardly be called a fan of Apple’s typical thermal solutions) explains here:

just to add here. Right after the part about sensors he states that ramping the fans to max should "lock" max performance of the chip. Obviously that is due to lower temps.
Literally what I am saying here.
Add a minute here and there and by the time the computer is ready to be replaced you made that extra $$ just from having a machine run always at 100%.

It's not just about longevity or chip/parts failure it is also about keeping peak performance as uninterrupted as possible.
To me this is worth million times more than noise or potential extra fan wear across computers lifespan. Easier to replace a fan than trying to make up for lost business.

 

[IG88]

You need to notice the parts where they mention their study is based on continuous operation at temperature, and 10 year lifespan at 105C is a figure they think is reasonable.

Do you use your laptop computers at maximum load 24/7/365 for 10 years straight? No? Duty cycle matters.

Outside of that PDF, you need to notice that Apple doesn't even let temps reach 100C continuous. It may hit that number briefly, but the control loops seem tuned to reach equilibrium somewhere in the mid-90s.

Maybe Apple's engineers aren't quite as grossly incompetent as random armchair-expert forum posters like to believe.

I noticed it. This was his claim:

There is no empirical data to corroborate the myth of "high internal temperature is bad for internal components".

Clearly that statement is nonsense. His own link spells that out quite plainly.

And Apple has had plenty of heat related issues in MacBooks over the years, partly due to inadequate cooling. Even as recently as the 2018 i9 MBP and the 2019 MBA. I never presented myself as an expert. Just someone that knows a ******** statement when they see it.