High Cpu Temp On Z170 Only At Uefi Bios (6700k)

Hi.

I have the same issue with my Extreme 7+ (6700k + H110i GTX Cooler).
But my pc works just fine. I think its just some bug in UEFI. It doesn't show correct temp sensor readings.       

I have noted higher temps in the UEFI on my Skylake systems but nothing close to 60c. On my Fatal1ty E3V5 Performance Gaming/OC I see temps up to 50c in the UEFI with my 6600k. I am using an Arctic Freezer 12 on that system. Once in windows however my temps dropped to 24c idle and about 62c under full load, Prime 95 + Furmark running on the iGPU. With a discrete GPU and just prime 95 I see about 55c max in OS. 

Thanks for the info. By the sound of it the BIOS is not reading the sensor correctly. If full fans doesn't impact it at all then it is seemingly fixed at 60c. It may be an inherent bug with the 6700k in particular because as Parsec noted his (and my own) 6600k displays more reasonable temps in UEFI.

erezc24 wrote: my cpu cooler is the zalman cnps x10 performa which works perfectly on other 1151 z170 boards like ASUS and Gigabyte  (46c idle) i have the  thermaltake mozart tx with 7 120" fans and im not connecting them to cha fan , i dont need to monitor them and im not using any custom profile with no OC   cpu cooler is connected to cpu fan1 of course and  even when i choose full speed the temp remain the same + the room is well air conditioned.   vcore voltage and status are fine as well and it seems that only the 6700k having trouble   im sure im not the only one with that issue, even at youtube  ppl wondering about that....   https://www.youtube.com/watch?v=AZpYXlDe3Uw   https://www.youtube.com/watch?v=ExDuTXE9atg   https://www.youtube.com/watch?v=HKdkU5aDras       i think those posts are more then enough and asrock should release some kind of fix or bios update because it cant be something else , like i said iv'e tried different z170 asrock models with the 6700k and all of them with same issue .

I'm sorry but your statements lack the detail that we need to understand what is going on. Without that detail, and more information, no mother board manufacture can take your claim seriously. I'm not saying absolutely you are wrong (clear the UEFI/BIOS lately?) just a reality check.

I also must say you seem to have completely ignored my explanation about why the CPU temperature is higher in the UEFI/BIOS.

Ok, you are using the ZALMAN CNPS10X Performa, a standard single tower cooler, five heat pipes, one 120mm fan, 900 - 2,000RPM without the quiet mode fan speed reducer, which results in 900 - 1,350RPM when used.

"Choosing full speed" for your CPU Cooler, on the CPU_FAN1 header and the CPU temperature remains the same? What does that normally mean when seen in a forum post about a CPU temperature issue?

A problem with the CPU cooler (fan speed), its mounting and contact with the CPU's IHS, use of too much TIM.

No CPU cooler fan speed readings in the UEFI or OS, no VCore or VID readings in the UEFI or Windows? You simply say they are fine, someone else can say they aren't, who is right?

Have you ever considered that your mounting of your CPU cooler was not perfect this time? I know the one I just did on the PC in the picture above is not perfect, since the fourth core is higher than it should be.

On the other hand, the i7-6700Ks have been a bit weird from the start. Super high VIDs for stock clocks (~1.4V). A strange shortage of them at retailers in the USA when first released. If you've seen this situation with several ASRock boards and UEFI versions for those boards, what is the one variable that has not changed? Your CPU. That's another reason I am skeptical about this.

Do you auto-run A-Tuning in Windows... why do I keep asking.

Xaltar, do you recall DooRules having a CPU temperature issue in the UEFI with his i7-6700K? I don't.

Another thing which I forgot to mention, and a question for those in this thread: What is the single CPU temperature reading, compared with or relative to the four core temperature readings we can find in many hardware monitoring programs that runs in Windows?

Eric's post about his CPU temperature in the UEFI and the core temperatures in Windows is a perfect example of the difference between these two things. Here's an example from an older PC I'm randomly using today, my ASRock Z77 Extreme4 with an i5-3570K:



At the bottom left is a program that shows which C State the CPU cores are operating in. Note I'm in 100% C0 state when the screen shot was taken, so no CPU power saving options active. Earlier, while I had the same instance of the HWiNFO64 monitoring program running, with all the C States enabled, you can see my minimum core temperatures. 16° C for Core 0 is pretty crazy low. This is using a Scythe Mugen II CPU cooler with modified mounting hardware.

We can see the difference in core temperatures with C States enabled and disabled. Over twice the temperature for some cores with C States disabled.

HWiNFO64 also displays the CPU Package Temperature and the CPU temperature from the PECI, the Intel Platform Environment Control Interface to the CPU. PECI uses a direct connection to the DTS temperature readings in the CPU, and uses the processor's internal PROCHOT value for temperature calculation, so is not affected by modifications to that reference value that can be done in the UEFI or OS.

We can see the two single CPU temperatures and the individual core temperatures are all somewhat different. So what do we see in the UEFI? Most likely the CPU (PECI) reading.

By the way Eric, I use a Noctua NH-D14 CPU cooler with the same CPU and mother board as you do. I also get sub-ambient temperature dips in my core temperatures at idle when using the CPU power saving features.

Have you ever had a piece of metal like a wrench sitting on a wooden desk or workbench, inside your home? Touch the wooden desk, it's not cold. Touch the wrench and... it's cooler, isn't it? It feels cooler, since metal is a great conductor of heat, which is why it is used in heat sinks. I have no problem believing momentary, small sub-ambient core temperatures, I used to see them on my i7-2600K which only had the C6 C State. Skylake has C7, which shuts down a core even farther. Plus the accuracy of the core temperature readings are less accurate at lower temperatures, according to Intel.

Notice now that we have one vote from Eric that the temperature in the UEFI is correct... oops, using an i5-6600K.

Finally, I recall a similar "issue" with an Asus P67 board I had. Owners complained about higher CPU temperatures in the UEFI/BIOS compared to Windows. This was a board used with the first Sandy Bridge processor generation released by Intel, which was quite different than previous generations. Differences like the temperature reading in the UEFI/BIOS were to be expected, such as which temperature value was displayed in the UEFI/BIOS. Owners just kept complaining.

So after a UEFI/BIOS update for that board, I noticed the CPU temperature in the UEFI went down 10° C compared to the previous UEFI version. Not about 10° C, exactly 10° C. I could load the previous UEFI version again, and the CPU temperature in the UEFI went up 10° C. I was getting CPU temperatures in the UEFI in the low 20s C with the new UEFI version.

The funniest thing was, I could see the adjusted CPU temperature using HWiNFO64. I had a 14° C CPU temperature reading, but also a correct reading of 24° C. Changing the UEFI version would cause that temperature reading to change in HWiNFO64 too.

I created a support ticket about this, but never received any reply. Meanwhile, the complainers were happy and stopped complaining.

After that, I bought my first ASRock board.

milchuck wrote: Recently built computer, pairing AsRock Extreme 7 with 6700K and Noctua NH-D15S.  Everything runs well and the cooler is huge but have the same unusual CPU temp readings....UEFI shows CPU as 55-56C while Core Temp and Real Temp both show around 26-28C idle.

This is explained in detail above, mainly in the post directly before yours.

CPU temperature is always higher in the UEFI/BIOS, than when in the OS/Windows. It's been that way with every PC I've ever built over the years. No one has ever noticed this until now?

What's your VCore reading in the same screen, H/W Monitor? Are you using default processor voltage settings?

What is actually unusual is the stock clock Intel VID for Skylake i7-6700K processors. A VID of 1.4V is common, and a VCore value close to that is to be expected. That is different than all earlier Intel processor generations, and much greater.

I guess Asus was right after all. Explaining why the CPU temperature is higher in the UEFI/BIOS does not make any difference. The temperature reading just needs to be lower.