What the sudden instability?

Hi guys,

I'm having trouble understanding how my CPU could be so unstable in such a short amount of time. When I first bought the 7700k, I cranked it up to 4.8ghz at 1.25volts and my memory at 3000mhz no problem. Worked great for many weeks...and now, I can't maintain a 4.8ghz speed at 1.36volts at 3000mhz speed without blue screening every time. What gives? If i drop the memory speed down to a default of 2133, I won't crash. But I want to run the speed at which I bought the ram in the first place. How can, within a matter of weeks, the stability of my system be such garbage compared to when I got the components? My temperatures are great. I never exceed 68-69 degrees so that isn't the issue.

Does anyone has any insight on this? It's very discouraging to have such a massive change in a short amount of time.

Components:

7700k
MSI z270 Gaming m5
Corsair 3000mhz 16gb
EVGA GTX 1080
EVGA 750w P2
11 answers Last reply
More about sudden instability
  1. is that a set voltage or an actual voltage?
  2. It's the voltage that I manually set within BIOS. That's another thing though. Whenever I set "AUTO" to voltage in the BIOS, I still crash. How can setting the voltage to AUTO not be enough for the cpu/mobo to determine what is required for me to be able to run the settings I have chosen.
  3. what's your actual voltage before the instability and after? not just the set voltage. What's LLC set at?

    I suggest you start again probing what's stable at a reasonable voltage.
  4. At default/turbo boost speeds, and the voltage is set to auto in bios, I'd say its anywhere between 1.2-1.25 volts. As I stated above, I was able to achieve a 4.8ghz at 1.25-1.275 volts for weeks when I first got the cpu. Now, i cant at 1.36 volts and 3000 memory speed. I've not had any experience with LLC so that setting is just set to auto as well.

    I've started over and probed many times at this point, why would the voltage need to be turned up so much higher than when I first got the cpu? And why now is it so unstable when I turn my ram speed to 3000mhz. None of these were issues in the first few weeks when I got the cpu. I ran an OCCT test for 10hours at 4.8ghz at 1.30 volts and it passed.
  5. well there are two options. Either your mobo started out supplying the voltage you needed and now doesn't i.e. it's drooping more, or the CPU now needs more voltage to be stable. (or it wasn't as stable as you thought).

    You voltage should be safe enough, therefore it must be that the mobo is not supplying as well as it once was.

    However without actual readings of actual voltages, not set voltages, then we won't know.

    did you research OCing first? LLC would be covered. Have you read this? http://www.tomshardware.co.uk/forum/id-1800828/intel-temperature-guide.html

    Personally I think that 4.8 is a stretch, and would be very sensitive to voltage.
  6. 13thmonkey said:
    well there are two options. Either your mobo started out supplying the voltage you needed and now doesn't i.e. it's drooping more, or the CPU now needs more voltage to be stable. (or it wasn't as stable as you thought).

    You voltage should be safe enough, therefore it must be that the mobo is not supplying as well as it once was.

    However without actual readings of actual voltages, not set voltages, then we won't know.

    did you research OCing first? LLC would be covered. Have you read this? http://www.tomshardware.co.uk/forum/id-1800828/intel-temperature-guide.html

    Personally I think that 4.8 is a stretch, and would be very sensitive to voltage.


    So you see this as a motherboard issue more so than a cpu issue? IF SO, what would be the reasoning behind my motherboard all of a sudden not being able to give out the required voltage to keep the overclock? IF it is the cpu, same question. Could either one of these have been faulty from the start?

    And you don't think 4.8 is achievable?
  7. components age when they are stressed, I say less likely to CPU as the voltage wasn't too high. But your mobo, was that overly stressed?

    Bear in mind when you OC ram there are three things that you are running above their base clocks, the RAM (should be no problem), the mobo and it's traces etc. and the memory controller on the CPU, what voltage are you running the memory at?

    Back off the memory OC and see if the CPU OC is stable, then back off the CPU OC and test the memory OC.

    Have you read the OC guides?
  8. 13thmonkey said:
    components age when they are stressed, I say less likely to CPU as the voltage wasn't too high. But your mobo, was that overly stressed?

    Bear in mind when you OC ram there are three things that you are running above their base clocks, the RAM (should be no problem), the mobo and it's traces etc. and the memory controller on the CPU, what voltage are you running the memory at?

    Back off the memory OC and see if the CPU OC is stable, then back off the CPU OC and test the memory OC.

    Have you read the OC guides?


    I've done both actually. When overclocking the CPU and having the memory at 2133 speeds, I don't blue screen. When running the cpu at default speeds, I am able to run the memory at 3000. As for the voltage on the memory, doesn't xmp make sure to increase the voltage as needed for the desired speeds? When overclocking my cpu and ram together, I never increased the ram voltage.

    And yes I have read the guides. I'll have to play with some of that when I get home.
  9. so what if you do memory at 3000 and CPU at 4.7.

    You need find that sweetspot again, as to why it's drifted... a bit of aging that might occur in the first weeks and months, the heatsinks over the VRMs could be slightly clogged making them run hotter, the air temp could have increased making the voltage droop slightly as the VRMs perform worse.
  10. 13thmonkey said:
    so what if you do memory at 3000 and CPU at 4.7.

    You need find that sweetspot again, as to why it's drifted... a bit of aging that might occur in the first weeks and months, the heatsinks over the VRMs could be slightly clogged making them run hotter, the air temp could have increased making the voltage droop slightly as the VRMs perform worse.


    Aging can occur within a few weeks time? That is insane.
    Going from 4.8ghz 1.25volts and 3000 memory speed to 4.8ghz 1.36volts and blue screening at 3000 memory speed just doesn't sound like an aging problem in this particular case though.

    I can certainly try that speed that you recommended though
  11. electronics age a little rapidly and then stabilise for a long time, then age again after that long time. You just have to find the spot again.
Ask a new question

Read More

Overclocking Speed Memory