Temp on i7 8700k

[rvxspeed]

Are these CPU core temps normal for 8700k cooled by Corsair H110iGT with an ambient temp of 24.8*C. I have noticed despite high temp CPU radiator is not so hot to touch. It looks like core temperature is rising regardless of being watercooled & not being transferred properly. Pump didn't failed because i can see pump rpm on corsair icue.

All other component used for reference:

Gigabyte Z370 Gaming 5
i7 8700k(just using turbo no OC) voltage set to 1.3 tried auto didn't work
EVGA RTX 2080
Samsung 860(250GB+500GB) nvme m.2
WD 6TB Black
Corsair H110iGT

Corsair AX1200i

Creative ZXR

[jj9987]

That's a whole lot of useless information in the post - we don't need your Windows, drive or network information here.

 

CPU load temps reach 83 - that's a bit on the high side.

[Quadrum]

Often times cpus can be much hotter than the radiator for a number of reasons, mostly bad thermal contact. From the images I’m not seeing anything too out of the ordinary. Is there something I’m missing?

[Mateyyy]

Have you updated your motherboard's BIOS? Those spikes to 1.46-1.48V are too high for the CPU, and it shouldn't even need that much voltage for 4.7GHz anyway.

[BobbyPdue]

I'm not sure what the issue is.  What are you running when your CPU goes up to 83c?  Has temps always been that high or is this the first time you've noticed them?  

[.Apex.]

47 minutes ago, jj9987 said:

That's a whole lot of useless information in the post - we don't need your Windows, drive or network information here.

Speak for yourself, this is very useful especially when someone is not sure which sensor to follow, lots of people use HWmonitor which is inaccurate as hell or show VID voltages which are useless.

 

54 minutes ago, rvxspeed said:

Are these CPU core temps normal for 8700k cooled by Corsair H110iGT with an ambient temp of 24.8*C. I have noticed despite high temp CPU radiator is not so hot to touch. It looks like core temperature is rising regardless of being watercooled & not being transferred properly. Pump didn't failed because i can see pump rpm on corsair icue.

I'd say this is pretty normal considering the voltages the CPU is running at are hovering around 1.35-1.45V, you could be running at 5Ghz comfortably with that voltage, if you want to lower your temperatures at 4.7Ghz you gotta do manual voltage tweaking which should be around 1.2-1.3V, I'm not sure why it's spiking that high since you've set it to 1.3V but I'm not familiar with Gigabyte boards, it'll probably help a ton to look up an overclocking tutorial to set the voltage properly and look into adjusting your LLC (Load Line Calibration).

[Blademaster91]

The Vcore around 1.45v can cause the cpu to run more hot, 83 is a bit high but still fine for 100% load.

[rvxspeed]

1 hour ago, jj9987 said:

That's a whole lot of useless information in the post - we don't need your Windows, drive or network information here.

 

CPU load temps reach 83 - that's a bit on the high side.

Sorry. I gave those information to create a reference compare to other component of the system.

[rvxspeed]

1 hour ago, Mateyyy said:

Have you updated your motherboard's BIOS? Those spikes to 1.46-1.48V are too high for the CPU, and it shouldn't even need that much voltage for 4.7GHz anyway.

Yes i did bios update recently. Couldn't recall the changes of temperature because my system had memory swapped recently in a very ackward way.

[rvxspeed]

1 hour ago, BobbyPdue said:

I'm not sure what the issue is.  What are you running when your CPU goes up to 83c?  Has temps always been that high or is this the first time you've noticed them?  

This is using a combined run of the application prepard3d & deliver us the moon. This is after recently i have made an upgrade to ram.

[rvxspeed]

50 minutes ago, _Syn_ said:

Speak for yourself, this is very useful especially when someone is not sure which sensor to follow, lots of people use HWmonitor which is inaccurate as hell or show VID voltages which are useless.

 

I'd say this is pretty normal considering the voltages the CPU is running at are hovering around 1.35-1.45V, you could be running at 5Ghz comfortably with that voltage, if you want to lower your temperatures at 4.7Ghz you gotta do manual voltage tweaking which should be around 1.2-1.3V, I'm not sure why it's spiking that high since you've set it to 1.3V but I'm not familiar with Gigabyte boards, it'll probably help a ton to look up an overclocking tutorial to set the voltage properly and look into adjusting your LLC (Load Line Calibration).

I have tried to booth using 1.3v. It boots then crashed inside windows. That's why i set it to auto.

[rvxspeed]

Ok guys. At first thank you so much for your responses. Here is a little bit of history to the problem. I originally built the system using 16GB 2x8GB vengance rgb kit from corsair. Later i decided to make an upgrade & go to 32GB using the existing 16GB kit & later just added the ram using the same sku. Turns out that my system wouldn't boot with xmp enabled. But, after i disabled it, the system ran for a month then the new 16GB kit ram gave up. It wouldn't even boot with a single stick of the new kit. So, i sent it for an rma. Corsair did a great job & gave me an upgraded version of the kit the vengance rgb pro. After recieving the product i had no problem on booting the system. So, it turns out that it was the new kit(which i sent for rma) that was faulty & after a month it gave up. Now the thing is with 16GB i was able to run 4.7(turbo,not oc) on all cores @1.2V. But, with 32GB kit even 1.3v isn't enough. So, i just set it to auto. Doing so raises the core temps significally. Also the core temps jumps around too much. My question is despite having a liquid cooler isn't that a bit awkward that it reaches above 80*C & not to feel warm to touch physically to the radiator itself. To point the validaty of my suspicisios i couldn't even hold my finger on my evga RTX 2080 radiator (120mm) more than 5 secs at full load where it shows 50*C core temperature.

[Mick Naughty]

Voltage is way to high for those low clocks. 

[Neggy-Z]

Unless you got really unlucky in the silicon lottery your vcore voltage does seem a bit high.

Since your running 4 sticks of memory you should try increasing the memory, vccio and vccsa voltages.

Set memory voltage to 1.35v, set both vccsa and vccio to 1.25v.

After that try lowering your vcore voltage.

 

[Tweedilderp]

Yeah do not go with a dynamic or "auto" vcore voltage.  That will degrade your silicon in no time.

 

I suggest keeping vcore at 1.32v max if you want tame temperatures.  I always manually set my vcore first to what I think is acceptable in regards to the die tolerances set out by intel and cross-reference with well-known overclockers on cpu owners' forums.  I myself have also delidded my 8700k and applied liquid metal, Also tried kryonaut (thermal grizzly) and found the temperature difference to be about 7c on average between the 2 but there could be margin of error in regards to IHS spacing in regards to application of the thermal compounds.

 

But your temps don't seem too bad considering your MOBO went nuts with the voltages so I would highly suggest adjusting to a static vcore of 1.32v max, if temps are fine with the clocks you get  you may increase clocks or voltage (if you decide to manually overclock).  I find that most 8700k users find the sweet spot is 4.7-4.9ghz, I run at 4.9ghz myself at 1.32v with a load line calibration to keep it 1.32v under load.

 

Never go by the VID voltage at the top of HWINFO as thats what the CPU is requesting, and like a Tweedilderp with a credit card in a PC hardware store, will never know what it should be settling for.  

 

If in the end you find that temps are tolerable and you have the clocks you want then...Success!, if temps are still not where you want them while staying under 1.35v on the vcore then that's when I may suggest looking for a local business or enthusiast with experience in delidding unless you are confidant you can do it yourself.

[rvxspeed]

10 hours ago, Tweedilderp said:

Yeah do not go with a dynamic or "auto" vcore voltage.  That will degrade your silicon in no time.

 

I suggest keeping vcore at 1.32v max if you want tame temperatures.  I always manually set my vcore first to what I think is acceptable in regards to the die tolerances set out by intel and cross-reference with well-known overclockers on cpu owners' forums.  I myself have also delidded my 8700k and applied liquid metal, Also tried kryonaut (thermal grizzly) and found the temperature difference to be about 7c on average between the 2 but there could be margin of error in regards to IHS spacing in regards to application of the thermal compounds.

 

But your temps don't seem too bad considering your MOBO went nuts with the voltages so I would highly suggest adjusting to a static vcore of 1.32v max, if temps are fine with the clocks you get  you may increase clocks or voltage (if you decide to manually overclock).  I find that most 8700k users find the sweet spot is 4.7-4.9ghz, I run at 4.9ghz myself at 1.32v with a load line calibration to keep it 1.32v under load.

 

Never go by the VID voltage at the top of HWINFO as thats what the CPU is requesting, and like a Tweedilderp with a credit card in a PC hardware store, will never know what it should be settling for.  

 

If in the end you find that temps are tolerable and you have the clocks you want then...Success!, if temps are still not where you want them while staying under 1.35v on the vcore then that's when I may suggest looking for a local business or enthusiast with experience in delidding unless you are confidant you can do it yourself.

Thanks for your valuable response. However i already tried with 1.3V didn't even boot. I am not planning to do overclock. Just 4.7 on all core will be more than enough. I will try with 1.35 or more on upcoming weekend. I am surprised of why my CPU is demanding such high vcore voltage with just an addition of 16GB ram to the system. Previously it was all good with 1.2v with 16GB kit. Even i have never seen temps go above 60*C.

[Neggy-Z]

11 minutes ago, rvxspeed said:

Thanks for your valuable response. However i already tried with 1.3V didn't even boot. I am not planning to do overclock. Just 4.7 on all core will be more than enough. I will try with 1.35 or more on upcoming weekend. I am surprised of why my CPU is demanding such high vcore voltage with just an addition of 16GB ram to the system. Previously it was all good with 1.2v with 16GB kit. Even i have never seen temps go above 60*C.

Populating all 4 dimm slots puts more stress on the IMC, if you have a weak IMC it will require more volts to maintain stability, check my post above.

[rvxspeed]

My system looks good. Doesn't it

[For Science!]

18 hours ago, rvxspeed said:

My system looks good. Doesn't it

It looks like a system only it's mother can love. I guess beauty is in the eye of the beholder

 

Spoiler

Maybe consider....

- Different photo angle, the view into the PSU shroud is not pretty

- Cleanup the PSU shroud area, send more cables to the back

- Route the GPU power cables over the gpu and into the rubber grommet under the 24 pin

- Clean the dust off the upper radiator.

- Match the color, not just the RGB values of the greens in the motherboard.

- Unless if you are going for a traffic light theme, do something about the red PSU, yellow fans, and green system

- You can set in BIOS to turn off the Q-code LED once booted.

 

On a more related note, if you would run RealBench instead of whatever you did, then I can say that with a 240 mm radiator in a very compact system I was running about 75 degrees with an all core turbo of 4.8 GHz (also auto voltage, on a Z370i Strix board). Delidding the chip reduced this by 15 degrees to reach 60 degrees instead. Manually overclocking and voltage adjustment later improved this further.

 

 

[rvxspeed]

3 hours ago, For Science! said:

It looks like a system only it's mother can love. I guess beauty is in the eye of the beholder

 

  Reveal hidden contents

Maybe consider....

- Different photo angle, the view into the PSU shroud is not pretty

- Cleanup the PSU shroud area, send more cables to the back

- Route the GPU power cables over the gpu and into the rubber grommet under the 24 pin

- Clean the dust off the upper radiator.

- Match the color, not just the RGB values of the greens in the motherboard.

- Unless if you are going for a traffic light theme, do something about the red PSU, yellow fans, and green system

- You can set in BIOS to turn off the Q-code LED once booted.

 

On a more related note, if you would run RealBench instead of whatever you did, then I can say that with a 240 mm radiator in a very compact system I was running about 75 degrees with an all core turbo of 4.8 GHz (also auto voltage, on a Z370i Strix board). Delidding the chip reduced this by 15 degrees to reach 60 degrees instead. Manually overclocking and voltage adjustment later improved this further.

 

 

Hi, thanks. I am well aware of delidding thing. But, it sounds dangerous. One small mistake & the CPU is dead. For now i will just stick to manual voltage control if i can get it to stable.

[rvxspeed]

here you go. although i am not a fan of asus.

[For Science!]

23 minutes ago, rvxspeed said:

here you go. although i am not a fan of asus.

Did you monitor per core temperatures during the stress test? Max of 88 is quite hot compared to 75 degrees for mine. Unless of course if your ambient temperature is 35+ degrees celcius.

 

Also good to make sure your mounting pressure is adequate by taking look at the thermal paste spread. 

 

[rvxspeed]

36 minutes ago, For Science! said:

Did you monitor per core temperatures during the stress test? Max of 88 is quite hot compared to 75 degrees for mine. Unless of course if your ambient temperature is 35+ degrees celcius.

 

Also good to make sure your mounting pressure is adequate by taking look at the thermal paste spread. 

 

No i didn't monitor invidual core. My ambient is 24~25 now-a-days. Mounting is good i believe. It was all good when i just had 16GB of ram. Now with 32GB it all running a bit toasty.

[For Science!]

6 minutes ago, rvxspeed said:

No i didn't monitor invidual core. My ambient is 24~25 now-a-days. Mounting is good i believe. It was all good when i just had 16GB of ram. Now with 32GB it all running a bit toasty.

Have you sanity checked your hypothesis? Pull out one stick of ram, disable XMP, set it at Vcore of 1.2 V and see if you can at least replicate your previous results.

 

Then stick the other stick in (still with xmp disabled) and see if you will boot. And finally enable XMP if previous two are okay.

[rvxspeed]

22 hours ago, For Science! said:

Have you sanity checked your hypothesis? Pull out one stick of ram, disable XMP, set it at Vcore of 1.2 V and see if you can at least replicate your previous results.

 

Then stick the other stick in (still with xmp disabled) and see if you will boot. And finally enable XMP if previous two are okay.

Only using 4 sticks raises the temps & doesn't get stable at 1.3v. Any other combination works just fine.