Which do I believe?

[OliverQueen]

I have noticed that the different software that I have which have the capability of monitoring temperatures etc all give vastly differing values to the point where I don't know which to believe!!

The Asus AI Suite for the ROG Maximus X Hero shows the lowest temperature values for the CPU by almost 20C at times ! The Corsair iCue software which I run to control the RGB on my peripheral devices also monitors the motherboard temperatures and shows CPU temperatures of around 30C at idle on all cores & under full load a maximum of 80C. HWINFO64 also gives the same readings give or take a couple of degrees either way to iCue.

As all 3 applications give differing values, which one is correct or most accurate??????? Both iCue & AI Suite run concurrently upon boot as AI Suite controls the fans & iCue controls the RGB on the RAM, keyboard, mouse, mouse pad. AIO fans, headset & headset stand. I have the RGB set to change depending on the temperature of the CPU but it goes "red" as set to 60C a lot of the time yet AI Suite still shows a maximum of 40C at the same time.

I obviously don't want to run the CPU too hot all the time as well as on the other hand mess around trying to get the values lower under iCue/HWINFO when they are the ones that are misreported!

i hate computers at time!!!!!!!!! LOL

 

[pendragon1]

what chip, what cooler?

 

[owcraftsman]

Although I prefer Aida64 Extreme paid version because of the OSD interface and customizable sensor panel Hwinfo 64 is what I would trust. I've used HWinfo64 for comparison sake vs Aida64 many times and they both tend to report similar results within a narrow margin. I would not trust iCue or AiSuite or generally any of the manufacturer software applications as they are 1/2 baked at best. Aida and HWinfo are dedicated to the one purpose only delivering harware monitoring and have perfected their craft over many years with constant updates.

 

[OliverQueen]

what chip, what cooler?

i7 8700 (non-K) with all cores running at a fixed 4.4GHz (set by XMP in BIOS) with a Corsair H115i Pro 280mm AIO with 4 x 140mm fans running Push/Pull. TIM is courtesy of a rice sized blob of CoolerMaster E1 "IC Essential" spread over the complete IHS surface rather than "squashed"

Although I prefer Aida64 Extreme paid version because of the OSD interface and customizable sensor panel Hwinfo 64 is what I would trust. I've used HWinfo64 for comparison sake vs Aida64 many times and they both tend to report similar results within a narrow margin. I would not trust iCue or AiSuite or generally any of the manufacturer software applications as they are 1/2 baked at best. Aida and HWinfo are dedicated to the one purpose only delivering harware monitoring and have perfected their craft over many years with constant updates.

I understand the theory etc but surely they are getting readings from the same diodes, how can the values vary so much? That is what I cannot get my head around. If the results were based on software calculations, I could understand the difference but as the temperatures are being provided by the same hardware, surely the result should be constant across all applications?

 

[owcraftsman]

The difference I already eluded to which is updates. The manufacturer's applications are rarely if ever are updated. As drivers, microcode, and the like are updated on your system; changes are sometimes needed to interpret data correctly. You probably know this but temps fluctuate rapidly, and polling rates and report rates are different with different software. Our processors, for example, raise and lower frequencies rapidly to conserve power when demand is low but the moment processing power is needed it climbs near instantly your voltages change as well, which cause heat and voltage spikes. Now take into account each of this application capture any given moment at different times and rates and reports them at different times and rates you can have striking differences as you have no doubt witnessed. This is another good reason to use good monitoring software that you trust. Things like polling and report rates are adjustable, but even if you start two different applications and match their rates with each other, they will still be capturing data at different points in time. To the best of my knowledge, there is no way to synchronize two different applications so in there lies the rub it will never match exactly the best you can hope for is using known good software that does it's a job well and use only one at a time to avoid confusion. I hope this makes sense.

 

[OliverQueen]

The difference I already eluded to which is updates. The manufacturer's applications are rarely if ever are updated. As drivers, microcode, and the like are updated on your system; changes are sometimes needed to interpret data correctly. You probably know this but temps fluctuate rapidly, and polling rates and report rates are different with different software. Our processors, for example, raise and lower frequencies rapidly to conserve power when demand is low but the moment processing power is needed it climbs near instantly your voltages change as well, which cause heat and voltage spikes. Now take into account each of this application capture any given moment at different times and rates and reports them at different times and rates you can have striking differences as you have no doubt witnessed. This is another good reason to use good monitoring software that you trust. Things like polling and report rates are adjustable, but even if you start two different applications and match their rates with each other, they will still be capturing data at different points in time. To the best of my knowledge, there is no way to synchronize two different applications so in there lies the rub it will never match exactly the best you can hope for is using known good software that does it's a job well and use only one at a time to avoid confusion. I hope this makes sense.

I understand completely and if t was a couple of degrees difference between all the pieces of software, I could accept it, but when you have both on screen at same time, and one is saying that the CPU Package is 35C and the other is saying that it is 55C without following a pattern which would suggest that polling rates are different, that is what is causing the confusion. I would have thought that the software for each of the manufacturers products would have been the most accurate due to being written specifically to operate with that hardware compared to a "one size fits all" like HWINFO or AIDA. The iCue software and Corsair hardware component firmwares are updated regularly (was an update to iCue waiting to be installed this morning). With the Asus software, all the Intel CPU's for each generation would all use the same monitoring components internally so the updates wouldn't really need changing unless a new chipset or microcode is released.

I suppose at the end of the day, if the rig is stable then it doesn't really matter which one is correct. I will probably just switch the RGB to cycle or stay static instead of changing depending on CPU temperature which would then be "out of sight, out of mind"

 

[owcraftsman]

I would bet if you stress test your system at 100% load all if not most of the applications would be reporting approx the same thing which is my point.

 

[doyll]

Exaxtly! Temp inaccuracies increase the farther away from critical temp we go .. meaning at 70c the error is like 3c but at 25c it can be 10c off.