Jul 26, 2019
21
0
10
Hey all,

I have done my first overclock ever of any CPU. I am able to get my 9900k to 5 ghz on all cores. (sometimes it hovers between 4998 and 5004 mhz) I have also set the ring frequency to 47, . Is this ok?
I wasn't sure how to do the voltage so I left it on auto to try to test the stability.
When running gaming and benchmark tests it looks like the voltage at the highest was hitting 1.305. Is this ok to leave on auto?
Will there be any negative aspects to leaving voltage on auto? I constantly have HW monitor open and I haven't seen it going any higher.
I am 100% new to overclocking CPUs and would welcome any insight you all have
I have a MSI MPG Z390 gaming pro carbon ac motherboard if that matters
EDIT: My temps have not exceeded 76 C

To recap, is 5.0 clock and 4.7 ring frequency goo,. or should I adjust the ring frequency?
Is it ok to leave on auto voltage in bios with 1.305 being the max volts I've seen?
Are there disadvantages to leaving it on auto voltage?




Thanks,
Andrew
 
Last edited:
Solution
Exactly. Run Prime95 with AVX/AVX2 disabled, which it gives you the option to do by unchecking the boxes next to one, then the other, and run for 15 minutes. If you exceed 80°C in HWinfo or Core Temp while running Prime then you need to either reduce the multi, reduce the voltage or get better cooling. A 360mm AIO is not comparable to a 360mm custom loop, by any stretch, but is still fairly decent if expectations are reasonable. There is a reason why they came out with a 9900ks and that reason is because an all core 5Ghz OC as daily driver is generally not reasonable on standard 9900k. Possible but not common, at least at that voltage and still be reasonably or fully stable.

For stability testing, as outlined in my guide, I'd recommend...
Ring frequency is ok, 45 might be better as a starting point though.

No, it is not ok to leave on auto. You will be leaving a lot of thermal headroom on the table by doing so. As a starting point, it is ok, but you don't want to leave it that way permanently.

I will tell you without ANY doubt whatsoever, that if you have a 5Ghz all core OC, with a core voltage of 1.305v, there is NO possible way you could have only a 76°C full load temp. Impossible. You'd have to be running a 360mm custom loop with terrific fans AND a very low ambient temperature in the room to do that. Maybe not even then. Not on that CPU.

My guess is that you have used a program or utility that does not in fact actually apply a 100% TDP steady state thermal load.

What is your CPU cooler model or type?

What utility or program did you use to stress the CPU in order to come up with that 76°C temperature?

How LONG did you run the stress test, if in fact you even RAN a stress test, to determine that your peak temp was 76°C?

Did you also test for stability, because testing for thermal compliance and testing for stability, are ENTIRELY different things and require entirely different processes.

I would HIGHLY recommend that you start by reading this:



And this:



.
 
  • Like
Reactions: CompuTronix
Jul 26, 2019
21
0
10
Ring frequency is ok, 45 might be better as a starting point though.

No, it is not ok to leave on auto. You will be leaving a lot of thermal headroom on the table by doing so. As a starting point, it is ok, but you don't want to leave it that way permanently.

I will tell you without ANY doubt whatsoever, that if you have a 5Ghz all core OC, with a core voltage of 1.305v, there is NO possible way you could have only a 76°C full load temp. Impossible. You'd have to be running a 360mm custom loop with terrific fans AND a very low ambient temperature in the room to do that. Maybe not even then. Not on that CPU.

My guess is that you have used a program or utility that does not in fact actually apply a 100% TDP steady state thermal load.

What is your CPU cooler model or type?

What utility or program did you use to stress the CPU in order to come up with that 76°C temperature?

How LONG did you run the stress test, if in fact you even RAN a stress test, to determine that your peak temp was 76°C?

Did you also test for stability, because testing for thermal compliance and testing for stability, are ENTIRELY different things and require entirely different processes.

I would HIGHLY recommend that you start by reading this:



And this:




Thanks for the response,
Again I am new and I appreciate the help. I just followed the guide for overclocking the 9900k from the MSI website for my mobo
I have a 360m AIO that I leave the fans on the highest setting.

As far as testing, I haven't really done a a true long term stress test. Just a few hours each of some CPU intense games. And a long run of 3d mark and superposition.

I will read these links you provided. thank you

What programs do you recommend for the testing?
8 hrs of Prime 95 and Realbench?
 
Jul 26, 2019
21
0
10
Thanks for the response,
Again I am new and I appreciate the help. I just followed the guide for overclocking the 9900k from the MSI website for my mobo
I have a 360m AIO that I leave the fans on the highest setting.

As far as testing, I haven't really done a a true long term stress test. Just a few hours each of some CPU intense games. And a long run of 3d mark and superposition.

I will read these links you provided. thank you

What programs do you recommend for the testing?
8 hrs of Prime 95 and Realbench?


I know I'm not as knowledgeable as you, but I have seen videos and posts where people are getting similar temps while gaming. That's all I will be doing with this CPU
 
Yeah, I'm skeptical of 5Ghz on all cores and only 1.305v too. I've got the same CPU. I can stick it at 5Ghz but it requires higher voltage than I am comfortable with and won't pass long term benching without clocking down. Instead I'm happy with 4800Mhz on all cores with a ring of 4600Mhz and voltage of 1.255.

Open up HWiNFO64 and leave in on while you bench and/or game. Then check min/max frequencies and 'Performance limit reasons'.
 
Exactly. Run Prime95 with AVX/AVX2 disabled, which it gives you the option to do by unchecking the boxes next to one, then the other, and run for 15 minutes. If you exceed 80°C in HWinfo or Core Temp while running Prime then you need to either reduce the multi, reduce the voltage or get better cooling. A 360mm AIO is not comparable to a 360mm custom loop, by any stretch, but is still fairly decent if expectations are reasonable. There is a reason why they came out with a 9900ks and that reason is because an all core 5Ghz OC as daily driver is generally not reasonable on standard 9900k. Possible but not common, at least at that voltage and still be reasonably or fully stable.

For stability testing, as outlined in my guide, I'd recommend you run Memtest86, run Prime95 with the custom options and run Realbench stress test. All tests times/lengths are outlined in there.
 
  • Like
Reactions: andrewcaldwell9
Solution
Jul 26, 2019
21
0
10
Exactly. Run Prime95 with AVX/AVX2 disabled, which it gives you the option to do by unchecking the boxes next to one, then the other, and run for 15 minutes. If you exceed 80°C in HWinfo or Core Temp while running Prime then you need to either reduce the multi, reduce the voltage or get better cooling. A 360mm AIO is not comparable to a 360mm custom loop, by any stretch, but is still fairly decent if expectations are reasonable. There is a reason why they came out with a 9900ks and that reason is because an all core 5Ghz OC as daily driver is generally not reasonable on standard 9900k. Possible but not common, at least at that voltage and still be reasonably or fully stable.

For stability testing, as outlined in my guide, I'd recommend you run Memtest86, run Prime95 with the custom options and run Realbench stress test. All tests times/lengths are outlined in there.

Thank you I will do all that and get back to you with the results.