Optimus V Extremely high CPU temperatures

C

carbide

Member
Hi all,

I have sent an email to the tech department at PCS, but I just thought I'd ask peoples opinions, is this normal?!

I just played a game for the first time this afternoon on my new Optimus V. was desperate to see the performance increase over my old laptop, and was pleasantly surprised to go from a 1366x768/med settings and 35 FPS to 1920x1080/high settings and 60 FPS - big thumbs up for the 765m!

I alt+tab'd out to check on CPU/GPU temps and was quite shocked to see core 0 had reached 97c and that there'd been a max temp differential of 16c across cores.

I downloaded OCCT so as I could put a controlled load across all cores and re measure as per the attached. The temp differential was lower at 10c as can be shown on the throttlestop max column.

I'm now using throttle stop to force a lower cpu state to avoid such high temperatures.

my question: are thermal sensors rubbish on haswell CPUs or is it more likely a bad TIM application?

my last laptop had a differential of 4c at any load state with standard paste, this new laptop I paid the extra for the arctic MX4 and i assume, hasn't pasted well?

GPU on the other hand is cool as a cucumber @68c with a fuzz test

thoughts?
 

Attachments

  • high temps 4700mq.PNG
    high temps 4700mq.PNG
    780.2 KB · Views: 222
C

carbide

Member
indeed!

well, I leave the laptop on the cooler, but with the fan to the cooler off. I don't want to affect the temperature results with the aid of a cooling solution external to the case until i'm happy that the laptop is capable of controlling it's own thermal environment. once i'm happy it is, i'll leave the cooler on, but until then i don't want to mask a cooling issue by trying to throw airflow at it. as the laptop is designed, it should be able to control the thermal demands placed upon it - indeed i'm surprised it didn't throttle by the time it got to 97c

i'm actually ok with that, as i can force lower the multiplier with throttlestop to a max temp i'm happy with, but it seems a shame to have to disable turbo boost completely if it turns out a poor TIM application is to blame rather than Clevo's chassis thermal design

I've just run intels processor diagnostic tool which doesn't give any info other than 'passed'. interestingly, it doesn't seem to be able to load the CPU anywhere near that of modern games!

the laptop fans run flat out when under full load - I've made sure there are no passive/active windows issues

regardless, I just don't think a temperature differential across a die surface so small is normal?! :-s
 
K

kylhan

Silver Level Poster
Don't worry!! Those temps are quite normal with the mobile i-7 processors believe it or not when under full load. Check this official response from Intel on the matter, hope this gives you some peace of mind

http://communities.intel.com/thread/29957 (This link will give you a certificate warning for some reason, but I've checked it and it's fine)

This is a copy of the reply from Intel on the matter if you don't want to click on the link.

The TCase for this processor is 105 degrees Celsius.
The TCase is a number established by Intel® as a point of reference in order to understand what could be expected as per normal processor temperature.

Anything from the Tcase and below will be the expected temperature of the processor in normal use, anything that doesn’t stress out the processor (watching movies, burning CDs, browsing the internet, creating documents, etc.) When the processor is stressed out meaning that you are running heavy processor applications that take control of the CPU or uses it at 100% the temperature will go beyond the Tcase. It can perfectly reach 115 to 120 degrees and the processor will still be OK. The cooling fan is in charge to keep that temperature there.

If the processor temperature reaches 130 degrees or more it will send a signal to the motherboard to shut down to prevent mayor damages and most likely it won’t be possible to turn the computer back in until it cools down.

The normal processor temperature will depend on the chassis type, the hardware involved and the location of the computer, and it usually is lower than the Tcase.

So the temperatures that you are reporting are OK, also I see that you are reading the cores temperature, you actually need to check the CPU temperature, and not the cores temperature.
In order to do so, please check with the laptop manufacturer to see if they have thermal monitoring software that can monitor the CPU temperature instead of the cores temperature.

Don't forget to +Rep if this helped
 
Last edited:
M

Matva

Member
Nice, if even Intel admits that temperatures up to 100 degrees Celsius are considered normal: No problem then, I guess :p It just looks so spooky because my desktop never goes over 60 and my previous laptop (a 17" gaming MSI) never got above 80.
 
M

Matrix510

Bronze Level Poster
My Optimus V 17" model doesn't go above 81 degrees while under full 100% stress test. Averages at about 78/79 degrees. I have an i7 4700MQ by the way, same as you.
 
C

carbide

Member
Phew, well, ok, where to start...

its been a long battle with pc specialist to try to educate that selling overheating, thermal throttling laptops is not a good business model. My suggestion that the temperatures are overly high has always been met with skepticism.

I myself was starting to believe that Clevo (the manufacturer of the laptops) was just a poorly designed budget brand offering max cpu configs way above their ability to cool. Still unconvinced, I carried out more testing:

https://dl.dropboxusercontent.com/u/44236923/Clevo W370ST full load temps.png

The image shows that three cores have hit TJmax, and the cpu has now initiated thermal throttling. I used to assume this was a semi aggressive multiplier reduction in steps, what it actually does is drop the multi to min (8x in my case) to dissipate the heat. Suddenly dropping down to an 800mhz cpu was the concerning behaviour seen in games which I was starting to think was a card issue.

Concerned, as much as I just wanted someone to build me a computer this time round, I got the heatsink assembly off.

CPU face - the two areas closest were actually completely dry, TIM had never come into contact with them
https://dl.dropboxusercontent.com/u/44236923/2013-10-31 20.32.22.jpg

GPU face - fairly even coverage - low load temps in operation
https://dl.dropboxusercontent.com/u/44236923/2013-10-31 20.32.40.jpg


So, If like me you're getting high core temps, or throttling, or, well, anything above mid/high 80's with the kind stress i'm producing in the photo above, you can be fairly certain there is something very, very wrong with the construction of you laptop. Also, if like me you paid £9 extra for the privlege of an overheating laptop, I would suggest after RMA you also request that back, seeing as I bought a tube to repaste with which cost the same and is good for another ten applications, I think that's ok!

I read a thread on this forum about overheating laptops when I ordered mine and it was in production, I was assured this was an issue of excess paste by inexperienced builders, make of that what you will. I also requested the hot room load temps whilst my laptop was in testing and was given 'CPU and GPU idle between 35-50, CPU peaks at about 65-70 and GPU peaks at about 75.' which obviously doesn't correlate at all to my testing, but perhaps I shouldn't have read too far into it - those could just have been generic 'should be' or 'expected' values

Additionally, after the QC process was complete and the laptop sent to me, it had the wrong ram installed, all three modules, so am still in the process of doing a swap with them.


On topic: after a repaste CPU idle temps have gone down from 55c to 40c, Load temps from TJmax (~100c) to 80c. The GPU, I thought the temps were ok and the paste job was ok but, after doing my own they've dropped from 68c to 60c and this is a slightly inferior paste.

on the upside, I managed to speak to a nice chap named Matthew Griffin who did accept PCS should have taken responsibility sooner and realised all was not fine, so that's something at least and has made me feel a little better. I'm just glad it isn't the case that Clevo have poor thermal management!
 
C

carbide

Member
kylhan said:
Don't worry!! Those temps are quite normal with the mobile i-7 processors believe it or not when under full load. Check this official response from Intel on the matter, hope this gives you some peace of mind

http://communities.intel.com/thread/29957 (This link will give you a certificate warning for some reason, but I've checked it and it's fine)

This is a copy of the reply from Intel on the matter if you don't want to click on the link.

The TCase for this processor is 105 degrees Celsius.
The TCase is a number established by Intel® as a point of reference in order to understand what could be expected as per normal processor temperature.

Anything from the Tcase and below will be the expected temperature of the processor in normal use, anything that doesn’t stress out the processor (watching movies, burning CDs, browsing the internet, creating documents, etc.) When the processor is stressed out meaning that you are running heavy processor applications that take control of the CPU or uses it at 100% the temperature will go beyond the Tcase. It can perfectly reach 115 to 120 degrees and the processor will still be OK. The cooling fan is in charge to keep that temperature there.

If the processor temperature reaches 130 degrees or more it will send a signal to the motherboard to shut down to prevent mayor damages and most likely it won’t be possible to turn the computer back in until it cools down.

The normal processor temperature will depend on the chassis type, the hardware involved and the location of the computer, and it usually is lower than the Tcase.

So the temperatures that you are reporting are OK, also I see that you are reading the cores temperature, you actually need to check the CPU temperature, and not the cores temperature.
In order to do so, please check with the laptop manufacturer to see if they have thermal monitoring software that can monitor the CPU temperature instead of the cores temperature.

Don't forget to +Rep if this helped
Click to expand...

Firstly, this isn't a dig at you, thanks for looking and finding info. I think it's really important to avoid aggregating the internet with false information - I don't know who 'Adolfo_Intel' is, it says he's a community member but I'm sure Intel wouldn't be too impressed to read his post on that page.

Tcase value of around 66-67c is a temperature related to Intel design documentation which is intended for the use of manufacturers when designing a laptop/pc's cooling system. The specific measurement of this value requires a channel to be cut into the IHS in order to sink a temp probe to lay flush with the IHS face so as a heatsink can be installed. There seems to be no info on the calibration procedures, type of sensor, value range etc. and regardless, it doesn't matter because it's of no use to an end user.

So forget about the existence of Tcase!

The 100c - 105c temperatures often mentioned refers to TJmax which signals the ceiling of the cpu's normal operation. At this point the CPU, upon reaching TJmax, the CPU initiates a hard throttle to force the frequency down and dissipate heat.

Approx. offset of 25c above TJmax, the CPU signals 'THERMTRIP' which, exactly as it sounds, is the final thermal shutdown temperature should throttling not be sufficient - I can only assume this comes into play if you do something daft like start the machine with no heatsink attached?!


So, you could argue that 99c is a safe stable operating temp, and we could just use mayonnaise as TIM. Refer to the Interesting article relating to current draw and thermal degradation over time:
http://www.anandtech.com/show/2468/6

It's very real! unsurprisingly, just like all other micro electronics, a high load, high temp environment will wear them out faster. Personally, my biggest concern is the m/board which has no way of regulating it's temperature.

I remember some of my fun pc builds with ridiculous CPU coolers like the scythe ninja? maybe which were so heavy, running high temps the entire m/board socket area sagged and visibly distorted!

kylhan said:
So the temperatures that you are reporting are OK, also I see that you are reading the cores temperature, you actually need to check the CPU temperature, and not the cores temperature.
In order to do so, please check with the laptop manufacturer to see if they have thermal monitoring software that can monitor the CPU temperature instead of the cores temperature.

Don't forget to +Rep if this helped
Click to expand...

No! the cores digital thermal sensors are calibrated at TJmax in order to give the CPU its most accurate reading as it approaches the TJmax temperature. There is nothing else, no other sensors placed so close to the point of heat output than the core DTSs. Forget anything else - any m/board thermal sensors are an approximation - how much is radiated heat? heat soak through the board etc.? I have no idea, but using a sensor at an unknown distance just monitoring its environment tells you nothing! I could run a stress benchmark, elevate the temps from idle to close to TJmax (like on the OEM original paste application of my laptop) within 30 seconds. Meanwhile the m/board sensor slowly moves a few degrees with considerable lag
 
You must log in or register to reply here.
Top