i5-8365U reaching 100ºC

[karmaikel]

The laptop looks a little warmer than in the reviews. If the bios locked to undervolt there are not many options... you can still change the thermal paste or lower the ambient temperature.

Such high temperatures will undoubtedly shorten the life of the laptop.

Actually afaik there's no review with my cpu so not sure if it is warmer than it should

 

[SL2]

Such high temperatures will undoubtedly shorten the life of the laptop.

No, not if it only stays there for a limited time like described. Those CPU's designed for high temps.

 

[karmaikel]

No, not if it only stays there for a limited time like described. Those CPU's designed for high temps.

How long that limited time should be? 10-15 secs max for example?

 

[repman244]

IMO this is normal, you could however try and replace the TIM, but since it's a business laptop that's up to you to decide.
You can also try and set maximum processor state (in windows power options) to 99% which disables turbo and try running like that just to see if there is any drastic change in temps.

 

[SL2]

Actually afaik there's no review with my cpu so not sure if it is warmer than it should

The first review I linked to is the same Whiskey Lake CPU as yours, but with higher clock speeds.

How long that limited time should be? 10-15 secs max for example?

Yeah I guess so, just like the review described.

 

[karmaikel]

The first review I linked to is the same Whiskey Lake CPU as yours, but with higher clock speeds.

Isn't it a Intel Core i7-8565U? i7... mine is i5, same number of cores and threads, don't know why one is i7 and the other i5 in the same generation

IMO this is normal, you could however try and replace the TIM, but since it's a business laptop that's up to you to decide.
You can also try and set maximum processor state (in windows power options) to 99% which disables turbo and try running like that just to see if there is any drastic change in temps.

Yes there is a drastic change in temps but also in performance

 

[SL2]

Isn't it a Intel Core i7-8565U? i7... mine is i5, same number of cores and threads, don't know why one is i7 and the other i5 in the same generation

In this situation the difference is clock speed and cache size, which is common for -U series. Hotter -H and -HK models are a whole different thing.

Intel product specifications

Intel® product specifications, features and compatibility quick reference guide and code name decoder. Compare products including processors, desktop boards, server products and networking products.

ark.intel.com

 

[Max(IT)]

No, it has no gpu

The 80º in a 40% load for example was executing this vbs script

while true
wend

it just goes up very fast, even with no demanding tasks. I guess it is its design. At the icon tray there is a tool where you can select 3 profiles, battery saver, balanced and performance. All my tests are on the balanced profile. With performance it doesn't go below base frequencies at idle, I don't like that.

Thanks for the answers. When I started this thread I thought that 100º was a dangerous temperature, 1º near some kind of 'meltdown', but @Max(IT) said emergency shutdowns are at 125/130C so I'm more or less safe. And the laptop is not even mine so I shouldn't care about it so much but I like things work properly and curious but the thermal configurations (my desktop pc never goes beyond 55C while gaming)

Now I know if one day a buy a laptop I will be very careful about its thermals...

Well you really cant compare a notebook to a desktop. The cooling system is much bigger and the airflow is different.

Unfortunately you cant say anything about the thermals of a laptop until you try it. I read some review with big differences in numbers.

Such high temperatures will undoubtedly shorten the life of the laptop.

Yes... from 10 to 9 years...

Nope, those temperatures are not going to reduce anything.

How long that limited time should be? 10-15 secs max for example?

Hours, not seconds.

 

[Valantar]

Isn't this just how laptops are designed these days? Ignoring core temperature and letting the chip bounce off TjMax, effectively self-regulating unless device skin temperatures go too high or power targets are exceeded? That's the only way to achieve the golden combination of a thin device with high performance and (relatively) low noise, after all.

Oh, btw, those of you saying devices are throttling because they are sustaining clocks far below boost? Please stop. By definition, modern CPUs are not throttling unless they go below base clock. Boost is never intended to be seen as a sustainable speed, and are typically given dramatically increased temporary power limits to achieve those speeds. I can agree that it's a slightly shady marketing tack (especially as base clocks are rarely marketed), but it is also a well established industry standard practice. Also, reality check: More than 4GHz across four cores at 15W? Are you insane? Even the more efficient Zen 2 on the more efficient TSMC 7nm process needs more than 3W per core at around 3.3-3.4GHz in a well-binned chip. 4GHz requires a lot more than that. Base clock is the rated sustainable clock speed when constrained by TDP or long term power limits, and while it is possible there is headroom for some sustained boost within TDP, it can never be expected. If you are reading a U-series (or even H-series) CPU spec sheet and you think the boost clock can be sustained beyond short bursts, you are deluding yourself.

I would say this laptop is likely working as intended by the manufacturer. It definitely isn't ideal, but it won't hurt the chip (though long term PCB damage or other thermal wear is more likely, many years down the line). The best you'll likely be able to do is tweak it so that it doesn't boost as high for as long. Ideally you'd be able to boost fan speeds, though that is exceedingly rare. Disabling Turbo Boost is a brute-force solution that will likely help, but at the cost of a very noticeable amount of performance.

 

[AsRock]

Unfortunately thats just how the cookie crumbles... HP have taken a page of out Apples book and put too powerful of a CPU inside a chassis that doesnt have good airflow or cooling to deal with satisfactory level.

I compared your 840 G6 with the newer macbook air (2020) running a slightly weaker CPU (i5-1030NG7) and that CPU still throttled in the review. I wouldnt expect your laptop to be any different. They are both the thin ultra portable style of laptop.

HP & Apple arent the only people that have done this. So have Dell & probably a few more companies. In their eyes, so long as the laptop doesnt crash because its getting too hot - its within operational parameters which means they wont accept it back for RMA. Even if the CPU is getting so hot its thermal throttling - Nobody will accept it back for RMA unless the laptop is completely dead or in a state where its not really fit for purpose.


Sadly this is just one of the nasty practises that laptop companies have picked up. They dont want you to have a well made laptop that lasts till the end of time. They want your new laptop to die within a year so you fork out money for another one. This has always been Apple's main modus operandi.

if their badly made laptops keep you coming back to them and paying for repairs - they get paid more money, That is how they do business.

::EDIT::

Alternatively, You could try repasting your CPU with a good thermal paste. The temps could drop anywhere between 5-8'c if not more.

The paste i replaced of our Dell machine it actualy had evaporated, iknow it had some on at some point as i had a repair guy out to iit and i was not to happy with how much that was put on it. Few years later i took a look in side and it all had pretty much gone lol.

 

[Valantar]

The paste i replaced of our Dell machine it actualy had evaporated, iknow it had some on at some point as i had a repair guy out to iit and i was not to happy with how much that was put on it. Few years later i took a look in side and it all had pretty much gone lol.

What? Thermal paste can't evaporate. The solvents keeping it soft can of course, so that it dries out, but the majority of the paste is made up of solid materials that would not evaporate unless exposed to a kiln or something equally blazing hot (at which point your laptop would of course also evaporate). Are you sure the paste hadn't dried out and crumbled as you disassembled it? Or that the repair guy didn't replace the stock paste with something else?

 

[Max(IT)]

Isn't this just how laptops are designed these days? Ignoring core temperature and letting the chip bounce off TjMax, effectively self-regulating unless device skin temperatures go too high or power targets are exceeded? That's the only way to achieve the golden combination of a thin device with high performance and (relatively) low noise, after all.

Oh, btw, those of you saying devices are throttling because they are sustaining clocks far below boost? Please stop. By definition, modern CPUs are not throttling unless they go below base clock. Boost is never intended to be seen as a sustainable speed, and are typically given dramatically increased temporary power limits to achieve those speeds. I can agree that it's a slightly shady marketing tack (especially as base clocks are rarely marketed), but it is also a well established industry standard practice. Also, reality check: More than 4GHz across four cores at 15W? Are you insane? Even the more efficient Zen 2 on the more efficient TSMC 7nm process needs more than 3W per core at around 3.3-3.4GHz in a well-binned chip. 4GHz requires a lot more than that. Base clock is the rated sustainable clock speed when constrained by TDP or long term power limits, and while it is possible there is headroom for some sustained boost within TDP, it can never be expected. If you are reading a U-series (or even H-series) CPU spec sheet and you think the boost clock can be sustained beyond short bursts, you are deluding yourself.

I would say this laptop is likely working as intended by the manufacturer. It definitely isn't ideal, but it won't hurt the chip (though long term PCB damage or other thermal wear is more likely, many years down the line). The best you'll likely be able to do is tweak it so that it doesn't boost as high for as long. Ideally you'd be able to boost fan speeds, though that is exceedingly rare. Disabling Turbo Boost is a brute-force solution that will likely help, but at the cost of a very noticeable amount of performance.

Correct.
My CPU (a more powerful i7-9750H) needs about 35W to stay all cores at 4 GHz, but sometimes it throttles back to 3.7 GHz. And there are spikes at 50W or above.

 

[AsRock]

Ok, to put it PC for you. about 10% of the cpu was still covered and all the paste from the whole tube he used had pretty much gone. Tell it shocking as i was expecting a big ol clean up for it which is partly why i left it a extra year.

If crumbled to dust i could not tell you but what was left was 100% useless.

 

[SL2]

Isn't this just how laptops are designed these days? Ignoring core temperature and letting the chip bounce off TjMax, effectively self-regulating unless device skin temperatures go too high or power targets are exceeded?

Well except that half of the thread is about how this is unlikely for this model. Elitebooks are about stability and reliability IMO, not most performance at any cost. I haven't seen any reviews of those that supports what you suggest.

Oh, btw, those of you saying devices are throttling because they are sustaining clocks far below boost? Please stop. By definition, modern CPUs are not throttling unless they go below base clock. Boost is never intended to be seen as a sustainable speed, and are typically given dramatically increased temporary power limits to achieve those speeds.

Thank you.

 

[Dante Uchiha]

Well you really cant compare a notebook to a desktop. The cooling system is much bigger and the airflow is different.

Unfortunately you cant say anything about the thermals of a laptop until you try it. I read some review with big differences in numbers.


Yes... from 10 to 9 years...

Nope, those temperatures are not going to reduce anything.

Nop. High temperatures and voltages accelerate the degradation of silicon, in addition to PCB and capacitors near the CPU. I've never seen a laptop that undergoes constant overheating last long. I had a macbook (core2duo) and an HP Elitebook(i7), I saw both die prematurely with just over a year.

 

[SL2]

I've never seen a laptop that undergoes constant overheating last long.

We're not talking about constant overheating here..

 

[Valantar]

Well except that half of the thread is about how this is unlikely for this model. Elitebooks are about stability and reliability IMO, not most performance at any cost. I haven't seen any reviews of those that supports what you suggest.

That's the thing though: these CPUs can handle this just fine. There is no reason for it to be unstable just because it is hitting TjMax and dropping to a lower boost level - 100C is entirely within the capabilities of a chip like that. And as I said this is not about getting the most performance at any cost, but rather about making a system that is simultaneously fast, slim and relatively quiet. It is a rather basic design principle of high end ultrabook-type laptops: they are typically too thin to cool their CPUs to low temperatures without running the fans very high, so something has to give. And seeing how the chips are running at safe voltages that can't be changed, low currents and moderate clock speeds, the solution becomes to let them run hot - with the addition of device skin temperature sensors to try to avoid this becoming uncomfortable.

 

[Max(IT)]

Nop. High temperatures and voltages accelerate the degradation of silicon, in addition to PCB and capacitors near the CPU. I've never seen a laptop that undergoes constant overheating last long. I had a macbook (core2duo) and an HP Elitebook(i7), I saw both die prematurely with just over a year.

that's just anecdotal evidence.
I know of many Macbook working for 10+ years without any issue at all.

 

[FreedomEclipse]

I know of many Macbook working for 10+ years without any issue at all.

This... is where i know youre lying Macbooks working for 10+ years?? Apple would like to know what sort of crack you've been smoking and where they can find these macbook owners so they can charge them for repairs and new macbooks they didnt buy as replacement for faulty (or just plain old) ones.

Apple would be out of business if their products lasted 10+ years. Nobody would be buying their stuff

 

[Max(IT)]

This... is where i know youre lying Macbooks working for 10+ years?? Apple would like to know what sort of crack you've been smoking and where they can find these macbook owners so they can charge them for repairs and new macbooks they didnt buy as replacement for faulty (or just plain old) ones.

Apple would be out of business if their products lasted 10+ years. Nobody would be buying their stuff

And this is where you just demonstrated you don't know what are you saying.
My old MacBook Pro 2009 is still working in my sister's house, with just an SSD and RAM upgrade over the years.

Said that, you were reported to moderators, because I'm not smoking any crack.

 

[FreedomEclipse]

And this is where you just demonstrated you don't know what are you saying.

My old MacBook Pro 2009 is still working in my sister's house, with just an SSD and RAM upgrade over the years.

I hate to break it to you but one person isnt 'many'

 

[Valantar]

This... is where i know youre lying Macbooks working for 10+ years?? Apple would like to know what sort of crack you've been smoking and where they can find these macbook owners so they can charge them for repairs and new macbooks they didnt buy as replacement for faulty (or just plain old) ones.

Apple would be out of business if their products lasted 10+ years. Nobody would be buying their stuff

Late 2000s/early 2010s MacBooks seem to be very reliable given the huge number of them still in use and on the used market. Later, hotter running models might be a different story (particularly with the garbage keyboard), but there is definitely some truth to older MacBooks being very reliable.

Now can we all please at least pretend we're adults here?

 

[Assimilator]

Tjunction for this particular CPU is 100 °C as per the Ark. Nothing to worry about, and nothing you can do to mitigate this with software like Throttlestop either, due to Intel's Plundervolt BIOS updates.

These temperatures are a function of cramming CPUs with ever-higher core counts and ever-higher frequencies, into ever-smaller chassis. If you want a CPU that won't thermal throttle in normal workloads, your only option is a desktop machine with adequate cooling.

You may be able to alleviate some of the throttling with a laptop stand with built-in fans, e.g. https://www.amazon.com/HV-F2056-15-6-17-Laptop-Cooler-Cooling/dp/B00NNMB3KS but that somewhat defeats the portability aspect of a laptop...

 

[AsRock]

Tjunction for this particular CPU is 100 °C as per the Ark. Nothing to worry about, and nothing you can do to mitigate this with software like Throttlestop either, due to Intel's Plundervolt BIOS updates.

These temperatures are a function of cramming CPUs with ever-higher core counts and ever-higher frequencies, into ever-smaller chassis. If you want a CPU that won't thermal throttle in normal workloads, your only option is a desktop machine with adequate cooling.

You may be able to alleviate some of the throttling with a laptop stand with built-in fans, e.g. https://www.amazon.com/HV-F2056-15-6-17-Laptop-Cooler-Cooling/dp/B00NNMB3KS but that somewhat defeats the portability aspect of a laptop...

Although i would check were the hot spots are and vents and get one with fans in the same kind of area's.

 

[95Viper]

Stay on topic.
Remember, that some people do not know if your expressions of speech are serious or not, because of regional differences in colloquialisms.

Thank you and have a nice day.