Question Hitting 94 C with i9 9900k and NZXT Kraken x72 on Stress Tests

Feb 12, 2019
9
0
10
Hey, I'm a bit new to the overclocking scene and I have OC'ed my i9 9900k to 5.0GHz. Running 3DMark (TimeSpy Extreme) I am hitting 90-94 on the CPU Stress portion of the test. I find this confusing as I have a 360mm Radiator, the Kraken x72. Is there any advice anyone can give about how to lower these temps while still having consistent high performance? Might I have a faulty part?

Thanks
 
May 4, 2019
59
6
45
Those temps are at the upper limit before throttling takes place, and you don't want to run there very long. The real way to diagnose these sorts of problems is to report the power use along with the temperatures. If you are running 250 watts plus, then the system is working as expected. If you are seeing those temperatures under 200 watts, then you have a cooling problem. Small ffts or other stress tests that crawl into the AVX/FPU unit and run entirely out of cache can draw a lot of power and generate high temperatures. Install CPUID HWmonitor or some other monitor software to get the whole picture of power vs. temperature.

My 9700k is unstable with the p95 small FFT test. It draws 250 watts, and with a 240 mm AIO, the temperatures run up to 100C at which point an error is generated. That's with a 5.0 GHz OC and MCE enabled. For my particular setup, 4.7 GHz is the highest OC I can run while using the small FFT/AVX2 stress test. At that point, the CPU is drawing about 220 Watts and runs at around 95C.

There is probably an AVX2 offset setting in the OC bios that will change the OC when the AVX2 unit is in use. I have not needed to use that so long as I stay away from the AVX/FPU stress tests, as normal use of the AVX2 unit runs just fine at 5.0 GHz.

The 9900k can draw a bit more power due to the additional hyperthreads, but I think that the AVX2/FPU stress tests will be similar. I don't remember offhand if the 9900k has additional AVX hardware over the 9700, but I would doubt it.
 

fagetti

Notable
Mar 1, 2018
919
15
1,165
Thats very high temperature, i always need to go way past any intel specs on the voltage ( with i7 cpu:s) to hit 90c on 3 fan aio cooling. I have corsair h150i 360mm aio which is slightly worse than yours in benchmarks and i need to get past 1.475v on gen1 i7 cpu:s to hit 90-95c.

What thermal paste you used? When you mounted the heatsink how much force you used to attach it with screwdriver?
 
May 4, 2019
59
6
45
Thats very high temperature, i always need to go way past any intel specs on the voltage ( with i7 cpu:s) to hit 90c on 3 fan aio cooling. I have corsair h150i 360mm aio which is slightly worse than yours in benchmarks and i need to get past 1.475v on gen1 i7 cpu:s to hit 90-95c.

What thermal paste you used? When you mounted the heatsink how much force you used to attach it with screwdriver?
Actually, those are expected temperature from this rig under the specified conditions. Without power measurements, you can't do a legitimate comparison. Temperature = ambient + power*thermal resistance. You said gen 1 cpu, which would be Nehalem, which does not support the AVX extensions. Those instructions are what the FFT stress test uses on 9th gen, so the anecdote is not particularly useful. Gen 1 may mean something different to you, as I have seen occasional articles with a different processor classification, but in either event, there will be fewer cores, and you also have a larger radiator.
 
  • Like
Reactions: fagetti

fagetti

Notable
Mar 1, 2018
919
15
1,165
Actually, those are expected temperature from this rig under the specified conditions. Without power measurements, you can't do a legitimate comparison. Temperature = ambient + power*thermal resistance. You said gen 1 cpu, which would be Nehalem, which does not support the AVX extensions. Those instructions are what the FFT stress test uses on 9th gen, so the anecdote is not particularly useful. Gen 1 may mean something different to you, as I have seen occasional articles with a different processor classification, but in either event, there will be fewer cores, and you also have a larger radiator.

Yeah i actually forgot all this AVX crap when stress testing, didnt know that it affects even 3dmark. Looking now i see that all of my gen1 cpu:s i7-875k i7-870 i7-970 etc dont have these avx extensions. I guess no way to overclock these i9:s with even kraken x72 its crazy how high temperature that is for just 3dmark (sorry for my bad english)
 
May 4, 2019
59
6
45
I'm not sure about the 3D benchmark, as I have not run it. But if that also winds up the onboard graphics, then it will also contribute to power draw. The gen 1 ran PCI 2.0, DDR3, and no onboard graphics, and SSE4(+)

As I said, give us the power v. temperature and we can then judge if the system is operating normally or has a problem.
 

CompuTronix

Intel Master
Moderator
Guys,

Utilities that don't overload or underload your processor will give you a valid thermal baseline. Here’s a comparison of utilities grouped as thermal and stability tests according to % of TDP, averaged across six processor Generations at stock settings rounded to the nearest 5%:



Although these tests range from 70% to 130% TDP workload, Windows Task Manager interprets every test as 100% CPU Utilization, which is processor resource activity, not actual workload. Core temperatures respond directly to Power consumption (Watts), which is driven by workload. Prime95 v29.8 Small FFT’s (AVX disabled) provides a steady 100% workload, even when TDP is exceeded by overclocking. If Core temperatures don't exceed 85°C, your CPU should run the most demanding real-world apps without overheating.

AVX - Advanced Vector Extension (AVX) Instruction Sets were introduced with Core i 2nd Generation CPU’s, then AVX2 with 4th Generation and AVX512 with later Generations of High End Desktop (HEDT) CPU’s as in certain X-Series, Extreme, i9’s and i7’s. Running versions of Prime95 with AVX enabled imposes an unrealistic workload which can adversely affect stability and severely overload your CPU to 130% TDP. 2nd and 3rd Generations are less affected, but Core temperatures on 4th through 9th Generations may be over 20°C higher.

Many 6th through 9th Generation motherboards address the AVX problem by providing “offset” adjustments (downclock) in BIOS. -3 (300 MHz) or more may be needed to limit Core temperatures to 85°C. Since 4th and 5th Generations don’t have AVX offsets, you can create individual BIOS Profiles for AVX and non-AVX software. Except for a few utilities and specialized computational apps, AVX intensive real-world apps (rendering, transcoding) and recent games with AVX shouldn’t exceed Prime95's test workload without AVX.

As per Intel’s Datasheets, TDP and Thermal Specifications are validated “without AVX”. In Prime95 versions from 27.7 through 29.4, AVX can be disabled by inserting CpuSupportsAVX=0 into the local.txt file, which appears in Prime95's folder after the first run. However, since Core temperatures will be the same as 29.8 without AVX, it's easier to just use 29.8. You can also use v26.6 which doesn't have AVX. Core i 1st Generation, Core 2, Pentium and Celeron processors don't have AVX Instruction Sets, so they're not affected.

See Page 87, Section 5.1.1 - Thermal Considerations - first paragraph, second sentence:

8th and 9th Generation Intel® Core™ Processor Families Datasheet, Volume 1 - https://www.intel.com/content/www/u...core/8th-gen-core-family-datasheet-vol-1.html

• Download the latest version of Prime95, which is 29.8 - https://www.mersenne.org/download/

This version finally provides an easy way to disable AVX, AVX2 and AVX 512 by just clicking on the boxes.

Also see Intel Temperature Guide - https://forums.tomshardware.com/threads/intel-temperature-guide.1488337/

You might want to read:

Section 11 - Thermal Test Basics
Section 12
- Thermal Test 100% Workload

CT 😎
 
Apr 2, 2019
45
8
35
I know all chips are different, but I think your cooler is not operating properly. I can run 5.4hz on my 9900k and finish blender runs with max temps of 85` with 240w loads on a 360 rad and custom loop. Have you tried setting a more aggressive profile for the cooler?