While stress testing a new CPU the temp went up to 100c, Any damage ?

[Davv]

So I overclocked a 9600K and then stress tested with Prime95 for a couple hours. A couple of cores reached 100c briefly a couple of times but mostly all cores hovered from upper 60's to the mid and upper 90's very often.

My CPU passed without any fails. I was wondering if I did any kinda of damage to my CPU cause of the high heat for those couple of hours ?

 

[luckymatt42]

No damage done. CPU has internal safeguards that will shut it down if temps get too high (unless you manually disabled those features in the BIOS). It was probably RIGHT on the edge of thermal shutdown, most intel will shut down around 100c.

 

[Davv]

J-temperature from Intel says 100c. I just really wanted that 5.0GHz overclock. In every day use including gaming temps will never even get close to what Prime95 was doing.

 

[luckymatt42]

That is absolutely true...but to ensure stability (and ensure no thermal limits are ever reached), you should still try to get that Prime95 down in the low 90's.

But having said that...which version of prime95 are you running? If it's newer than 26.6, the program is utilizing AVX instructions which does make the chip run hotter. But in any case, try to keep sustained loads max temps under 95 degrees.

 

[CompuTronix]

... overclocked a 9600K and then stress tested with Prime95 ... couple of cores reached 100c ...

Q: Which version Prime95?
You might have used the wrong version, which can cause very high and unrealistic Core temperatures. The correct version is 26.6.

Q: Which torture test?
Blend is a memory stability test. It has a fluctuating workload with fluctuating Core temperatures. Small FFT's is a CPU thermal test. It has a steady workload with steady Core temperatures.

Here's the nominal operating range for Core temperature:

Core temperatures above 85°C are not recommended.

Core temperatures below 80°C are ideal.

Core temperatures increase and decrease with Ambient temperature.

Q: What ambient (room) temperature?

Q: What cooler are you running?

Q: What Core voltage settings?

CT

 

[luckymatt42]

... overclocked a 9600K and then stress tested with Prime95 ... couple of cores reached 100c ...

Q: Which version Prime95?
You might have used the wrong version, which can cause very high and unrealistic Core temperatures. The correct version is 26.6.

Q: Which torture test?
Blend is a memory stability test. It has a fluctuating workload with fluctuating Core temperatures. Small FFT's is a CPU thermal test. It has a steady workload with steady Core temperatures.

Here's the nominal operating range for Core temperature:

Core temperatures above 85°C are not recommended.

Core temperatures below 80°C are ideal.

Core temperatures increase and decrease with Ambient temperature.

Q: What ambient (room) temperature?

Q: What cooler are you running?

Q: What Core voltage settings?

CT

(sorry for the detour, just have a question/comment)
I used to recommend 26.6 also, as almost no "everyday user" workloads used AVX. But I've seen recently where some applications (graphics rendering, etc) and now even games are starting to use the AVX instructions. So 26.6 is still somewhat realistic, but do you think using later versions with AVX is more relevant these days?

 

[CompuTronix]

luckymatt42,

In a word, no.

Real-world AVX apps, and the very few newer games that use it, are fluctuating workloads that don't saturate the processor with AVX codes, and thus, won't exceed the 100% workload of non-AVX Prime95 version 26.6. Utilities such as Prime 95 version 29.4 (AVX) will overload the processor to nearly 130% TDP workload, while Task Manager will still be reading 100% Utilization.

Intel tests their processors at a steady 100% TDP workload to validate Thermal Specifications. Utilities that don't overload or underload your processor will give you a valid thermal baseline. Here's a comparison according to % of TDP, averaged across six processor Generations at stock settings rounded to the nearest 5%:

All tests will show 100% CPU Utilization in Windows Task Manager, which indicates processor resource activity, not % TDP workload. Core temperatures respond directly to Power dissipation (Watts), which is driven by workload.

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 Generation don’t have AVX offsets, you can create a BIOS Profile for gaming, and a downclock Profile for AVX apps such as rendering or transcoding. If you don’t use AVX apps, BIOS should still be configured for it, as certain utilities use AVX for stability testing.

AVX can be disabled in Prime95 versions later than 26.6 by inserting "CpuSupportsAVX=0" into the "local.txt" file in Prime95's folder. However, since Core temperatures will be the same as 26.6, it's easier to just use 26.6.

As per Intel’s Datasheets, TDP and Thermal Specifications are validated “without AVX”.

8th and 9th Generation Datasheet, Volume 1 - https://www.intel.com/content/www/us/en/products/docs/processors/core/8th-gen-core-family-datasheet-vol-1.html

See page 87, Section 5.1.1 Thermal Considerations, first paragraph, second sentence.

You might also want to read this: Intel Temperature Guide - http://www.tomshardware.com/forum/id-1800828/intel-temperature-guide.html

See Section 11 - Thermal Test Basics, and Section 12 - Thermal Test 100% Workload

CT

 

[Davv]

I used Prime95 26.6 and my voltage was set to auto. During testing at 100% load it said that my voltage was at around 1.43 approx for 60mins with no failures. This was for a 5.0GHz overclock.

So are you saying that even at 100% load stress testing temps shouldn't be that high ? Then what would be the limit of the temps then, should I lower the overclock untill I my temps stay in the 80c and never into the 90c ? What's the highest temps should I allow even for stress testing ?

Last question what's the min amount of time do I need to stress test is one hour eoungh ? Please don't tell me that 12 hours, what's the min ? Room temperature by the way was 70 farenheit.

I did read a couple of parts of that guide it said above 85ﾟis not recommend, I'm assuming that's including during stress testing ?

Cooler I'm using is Noctau tower cooler with a 120mm fan.

I probably should of mentioned this earlier I'm using the Asus strix E z390 board and I enabled the AI optimizer where basically it overclocks the board for you. I know it's better to do it manually but I need a guide to tell me exactly what you do. This board has so many settings I just didn't want to fuss with it. Either I use the AI or I need a guide of sorts on which setting to change for a 5.0ghz overclock.

 

[Karadjgne]

TDP is also a little misleading. It's originally thermal design Power, as in the amount in watts that a cpu will use at 100% loads. Coincidence puts thermal watts generally within 5°C± of TDP, so figure at a 95w TDP cpu using @95w and giving off 95w of heat. This is important when choosing a cooler, but more so when it comes to OC as thermal watts will vary to greater degrees with higher OC, in some cpus the thermal watts can hit 2.5x TDP under OC conditions, even though electrical watts are not even close.

The i5-9600k is a 95w TDP cpu. At 5.0GHz expect it to hit 2x TDP in heat, so you'd be looking at using a cooler rated at @200w TDP in order to keep temps in check at @70°C or less. Use of a less effective cooler will still work, but under stress workloads like 80-100% you be looking at higher temps. Stock coolers are rated barely above the TDP of the cpu, so expect high workloads to hit thermal throttling quick with something like p95 being almost instantaneous throttling.

Will stress testing at 100°C damage a cpu? Unknown. Cpus are designed to first attempt throttling the speeds and voltages in an attempt to lower temps. If that finally fails, you get shutdown. A safety precaution. Ideally, there will be no damage. Repeated and forced usage at those temps is Russian Roulette, sometimes you get lucky, sometimes you don't. Damage is permanent.

Now knowing that you hit 100°C at 100% workload, expect temps to be somewhat similar to loads. This'll mean cpu punishing games at high % loads are going to show high temps.

I'd be looking at either dropping the OC or getting a better cooling solution.

 

[CompuTronix]

I used Prime95 26.6 and my voltage was set to auto ... voltage was at around 1.43 ... This was for a 5.0GHz overclock ... what would be the limit of the temps ... What's the highest temps should I allow even for stress testing ? ... what's the min amount of time do I need to stress test is one hour eoungh ?

There's the problem:

voltage was set to auto ... voltage was at around 1.43

Overclocking should not be attempted with Vcore settings in “Auto” because BIOS will apply significantly more voltage than is necessary to maintain stability, which increases Power (Watts) and heat (Core temperature).

1.43 volts is the extremely high, and exceeds the maximum recommended Core voltage, which is 1.400.

Here's the maximum recommended Core voltages per Microarchitecture from 14 to 65 nanometers since 2006:

When tweaking your processor near it's highest overclock, keep in mind that for an increase of 100 MHz, a corresponding increase of about 50 millivolts (0.050) is needed to maintain stability. If 70 millivolts (0.070) or more is needed for the next stable 100 MHz increase, it means your processor is overclocked beyond it's capability.

Remember to keep overclocking in perspective. For example, the difference between 4.5 GHz and 4.6 Ghz is less than 2.3%, which has no noticeable impact on overall system performance. It simply isn’t worth pushing your processor beyond recommended Core voltage and Core temperature limits just to squeeze out another 100 MHz.

There are plenty of good overclocking guides available. Research the topic before you attempt to overclock so you don't kill your processor. Here's a good place to start:

Beginners guide to overclocking your CPU (With explicit testing guidelines) - http://www.tomshardware.com/faq/id-3761568/beginners-guide-overclocking-cpu-explicit-testing-guidelines.html

... what would be the limit of the temps ... What's the highest temps should I allow even for stress testing ? ...

See the chart in my first post; it's quite clear.

... what's the min amount of time do I need to stress test is one hour eoungh ?

"Stress" tests vary widely, so any answers to that question depend on which stress test(s) you intend to use, and for what purpose. Stress tests can be characterized into two categories; stability tests which are fluctuating workloads, and thermal tests which are steady workloads.

For example, Prime95 v26.6 Small FFT's is better suited for "thermal" testing than any other utility. A 10 minute test is more than sufficient for air or AIO coolers. However, when Small FFT's is used for "stability" testing, an overnight run sometimes isn't long enough.

In contrast, Prime95's default test, "Blend", is well suited for testing memory stability, and should be run for about 2 hours, whereas Asus RealBench is great for testing overall system stability, and should be run for at least 1 hour. Opinions vary widely concerning how long to run which tests.

I strongly advise that you decrease Core voltage to less than 1.4. Your 9600K shouldn't need more than about 1.35 to achieve a stable 5.0GHz overclock, more or less. Core temperatures respond directly to Core voltage which drives power consumption (Watts). Lower Core voltage means lower Core temperatures.

CT

 

[Davv]

Can you guys read my last post, I've included much more to what is going and why I did certain things.

 

[CompuTronix]

Davv,

Overclocking is always limited by two factors; voltage and temperature. There are no guarantees you can hit 5.0GHz, because no two processors are identical; each is unique in voltage tolerance, thermal behavior and overclocking potential, which is often referred to as the "silicon lottery".

With high-end cooling you might reach the Vcore limit before 85°C. With low-end cooling you’ll reach 85°C before the Vcore limit. Regardless, whichever limit you reach first is where you should stop ... at that point, whatever overclock you get is what you get.

Although the majority of 9600K's can reach 5.0, some can only reach 4.8, while others can reach 5.3. But whatever overclock you can achieve, it should be within the maximum recommended limits of 1.400 Vcore and 85°C.

Personally, I don't like to see any Core temperature in the 80's ... ever!

Google "overclock 9600k asus z390" or similar. There are several guides available.

CT

 

[Karadjgne]

My 3570k will only hit 4.3GHz. Many others, most others hit at least 4.5 and don't top out until 4.8-4.9GHz. Seriously lost the silicon lottery with that cpu. My 3770k purrs like a kitten at 4.9GHz @1.32v. winner - winner chicken dinner.

It happens.

 

[Davv]

Davv,

Overclocking is always limited by two factors; voltage and temperature. There are no guarantees you can hit 5.0GHz, because no two processors are identical; each is unique in voltage tolerance, thermal behavior and overclocking potential, which is often referred to as the "silicon lottery".

With high-end cooling you might reach the Vcore limit before 85°C. With low-end cooling you’ll reach 85°C before the Vcore limit. Regardless, whichever limit you reach first is where you should stop ... at that point, whatever overclock you get is what you get.

Although the majority of 9600K's can reach 5.0, some can only reach 4.8, while others can reach 5.3. But whatever overclock you can achieve, it should be within the maximum recommended limits of 1.400 Vcore and 85°C.

Personally, I don't like to see any Core temperature in the 80's ... ever!

Google "overclock 9600k asus z390" or similar. There are several guides available.

CT

Update:

It's way better OC'ing manually then using a automatic settings, no doubt. So I was able to OC my 9600K to a stable 5GHz. I used 1.35 voltage and the max temp was 85c. I'm probably using too much voltage for a 5GHz OC.

Also whats weird is that in hwmonitor app it says that my max voltage was 1.244 during the stress testing at 100%. But I set the voltage manually at 1.35.

Even though I wont increase my voltage cause I'm already hitting 85c, Im thinking I may increase my OC to 5.1GHz and see if I can get it stable with 1.35. But im wondering if could it still increase my heat though ?

 

[Karadjgne]

Make sure you are reading what you think you saw. There's a marked difference between VID and vCore although many programs will only show one. My cpu-z msi version shows VID, but cpu-z ROG version shows vCore, and they are different.

VID is voltage identifier, it's what the cpu actually requests from the motherboard voltage regulatory circuitry and is a set figure. vCore is what the cpu actually receives from the mobo, and can change due to loads, demand, cpu, vDroop, power saving (eist etc). vCore as read by software can also be slightly inaccurate as it is read by software, and does change at any given moment, for any particular reason, depending on what the cpu is actually doing.