[SOLVED] 5.1 @ 1.275v or 5.2 @ 1.390v i9-10900k

[bishopi5]

Hey guys been screwing around with my new chip, bought the whole PC to overclock, have figured out i didn't really hit the silicon lottery with this one, although it runs at 5.1 on all cores at rather low voltage (1.275bios, about 1.266 load LL3)

I don't feel like this is enough of a boost and would much rather go down the road of buying a crazy cooler and running it at over 1.4 at 5.3 or something, im currently running nocuta nh-15

So my question is, considering the massive bump in voltage needed for it to run at 5.2, is it worth it? i have done a lot research and 1.390 volts seems pretty safe as long as temps are low and they are (in daily use)

Or should i settle for 5.1 @ 1.275v until i sell my noctua and find a nice cooler and go for 5.3 because i know thats stable at just over 1.415, any advice would be greatly appreciated

 

[jay32267]

"is it worth it? "
A 2% increase wouldn't be worth it to me

 

[hotaru.hino]

A 0.1-0.2GHz overclock at these speeds are a drop in the bucket and especially not worth it considering how much voltage you have to bump it up to. Voltage affects power dissipation exponentially.

So while a 0.1GHz bump by itself increases the power dissipation by about 2% here, bumping up the voltage increases it by nearly 19%.

 

[CompuTronix]

... runs at 5.1 on all cores at rather low voltage (1.275bios, about 1.266 load LL3) ... would much rather ... at over 1.4 at 5.3 or something ... nocuta nh-15 ... massive bump in voltage ... to run at 5.2 ... 1.390 ... should i settle for 5.1 @ 1.275v ... sell my noctua and find a nice cooler and go for 5.3 because i know that's stable at just over 1.415 ...

Your voltage and frequency values don't make sense ...

5.1 @ 1.275 to 5.2 @ 1.390 = 110 millivolts increase

5.2 @ 1.390 to 5.3 @ 1.415 = 25 millivolts increase

Here's an example of a Core Voltage / Frequency Curve:

With respect to the voltage / frequency curve, if your 10900K is indeed stable at 5.1 GHz with 1.275 Vcore, then 5.2 GHz should require no more than about 1.34 Vcore, whereas 5.3 could conceivably require as much as 1.41 Vcore to achieve stability. Your processor should follow a similar curve. Considering your Vcore value of 1.39 for 5.2 GHz is a 110 millivolt increase from 5.1, but only a 25 millivolt increase from 5.2 to 5.3, your numbers suggest that you may have been a bit hasty in finding your processor's stable overclock settings, and that your test methods were perhaps inconsistent and somewhat less than thorough.

To achieve the highest overclock, keep in mind that for your final 100 MHz increase, a corresponding increase in Core voltage 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 you're attempting to overclock your processor beyond its capability.

Each Microarchitecture has a “Maximum Recommended Vcore”. Here's the values per Microarchitecture from 14 to 65 nanometers since 2006:

Each Microarchitecture also has a "Degradation Curve". As a rule, CPUs are more susceptible to electromigration and degradation with each Die-shrink. However, the exception to the rule is 14 nanometer (nm) Microarchitecture, where advances in FinFET transistor technology have improved voltage tolerance.

Here's how the Degradation Curves correspond to Maximum Recommended Vcore for 14 nanometer 5th through 10th Generations, which for comparison, differs from 22 nanometer 3rd and 4th Generations:

It is typically recommended that you should not exceed 1.4 Vcore, however, if you look very closely at the chart, you can see that 1.425 is about as high as you should go on 14 nanometer processors. This agrees with the recommendations of Intel's own Overclocking Test Lab Engineers (see paragraphs 8, 9 & 10), who stipulate that you should not exceed 80°C. (See Intel CPU Temperature Guide 2021)

Nevertheless, although the Noctua NH-D15 is the top-of-the-line air cooler, it's not capable of adequately cooling the 10 Core 20 Thread i9-10900K at the Core voltages you're proposing, due to the extreme Power consumption (over 325 Watts) that such voltages will drive. Even at stock settings, a 360 mm AIO or a custom loop is instead recommended.

I suggest you heed the good advice that jay32267 and hotaru.hino offered, and read the links I provided to learn more about overclocking, so you don't inadvertently subject your new CPU to the effects of Electromigration and Degradation. It simply isn’t worth pushing your processor beyond recommended Core voltage and Core temperature limits just to squeeze out another 100 MHz, which is less than 2%, and has no noticeable impact on overall system performance.

CT

 

[bishopi5]

Thanks for the detailed anwer, and your are indeed right, i have no way of truly testing if 5.3 is 100% stable as it is only game stable at 1.415v, any game i throw at it.

Cinebench will almost certainly throw a BSOD if it dosen't thermal throttle before so

Yes it is indeed rock solid at 1.275v about 1.266 load voltage at LL3 at 5.1 all cores, (cinebench loops for hours, max temps around 81) does it make sense that i need 1.390v to run 5.2????

I am game stable at 1.355v but will blue screen if running less then 1.390v in cinebench r20

 

[CompuTronix]

CineBench stable 5.1 at 1.275 with maximum Core temperature of 81°C ... at what ambient (room) temperature?
CineBench stable 5.2 at 1.390 ... at what Core temperature?

Too much Vcore increase (115 millivolts) from 5.1 to 5.2.

Only game stable 5.3 at 1.410 ... what Core temperature?
Vcore is borderline too high.

A whopping 115 millivolts just to get from 5.1 to 5.2 should tell you that your particular i9-10900K sample is not capable of sustaining an all-Core long-term overclock at 5.2 GHz with reasonable Core temperatures. As I already pointed out, although the Noctua NH-D15 is the top-of-the-line air cooler, it's not capable of adequately cooling the 10 Core 20 Thread i9-10900K at the Core voltages you mentioned, due to the extreme Power consumption (over 325 Watts) that such voltages will drive. Even at stock settings, a 360 mm AIO or a custom loop is recommended.

While CineBench is a realistic rendering workload which is suitable for testing stability, it has very similar Power consumption and Core temperatures to Prime95 Small FFTs without AVX, which is a steady-state, true 100% workload that's ideal for testing thermal performance. Here’s a comparison of utilities grouped as thermal and stability tests according to % of TDP, averaged across six processor Generations at stock Intel 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 %TDP workload. Core temperatures respond directly to Power consumption (Watts), which is driven by workload.

If you expect to limit your 10900K to about 80°C, which is highly recommended, then your all-Core 5.1 GHz overclock at 1.275 Vcore is your best overclock to avoid the effects of Electromigration and Degradation. Once again, it simply isn’t worth pushing your processor beyond recommended Core voltage and Core temperature limits just to squeeze out another 100 MHz, which has no noticeable impact on overall system performance.

Please click on the red links above and read the information I've provided for you, which is as well for everyone's benefit.

CT

 

[bishopi5]

Ambient room temp is very low, around 19c and i have no side panel on, it might get a little higher then 81 but that seemed to be its cap

I appreciated your answers and will be sticking with 5.1 until i go with custom loop (always wanted to) but i had bad experience with new corsair closed loop that took out 2 of my sapphire vapor x 7970's, took months to figure out as the leak was that tiny is was more of a mist, and would dry out before leaving any residue... for a while anyways, been haunted by water ever since but i love overclocking and the 11900k was a fail so i can't see going to less cores in the future so maybe custom loop that i know i built will remedy any concerns i have of leakage.

But my main question was about my chip sample, it is indeed stable at 1.275 LLC3 5.1 all cores, fixed ratio.

From what you said it seems that it should only need 1.35ish to hit 5.2?? from what i can see 5.1 at such low voltage is pretty decent sample, yet it needs that big of a jump?

I can't test cinebench at 1.355+ because of my cooler all i know is 5.3 was game stable at 1.415 and 5.2 was game stable at 1.355-1.365, but 5.2 needed 1.390 to pass cinebench r20 and cores reached 100

UPDATE:

Tested 5.1 once again after researching your graphs and the links provided i was finding it very strange i needed so much of a bump to hit 5.2 (although was game stable for many days / different games)

5.1 threw errors in HWINFO, didnt BSOD or crash during cinebench r20 but threw some errors, i was gobsmacked, i know for a fact it was stable at 5.1 @ 1.275 LL3 (1,266 LOAD), the absolute only thing thats changed is my bios version, i updated it when trying to get 5.2 stable a while back, and it was very old

So im guessing my sample makes a bit more sense now, it needs 1.315v to throw no errors at 5.1, so 1.390v for 5.2 dosen't seem that far off

Thanks for your help