bmockeg :
@darkbreez, I will send you screenshots when I get home if you like. I ran AIDA 64 CPU for an hour no issues.
One hour of running Aida64 doesn't mean your system is stable. Aida itself recommends a minimum of 12 hours running their FPU stress test. And that's only a smaller part of the overall picture. There is, practically, no single test that will ensure stability.
When you have run multiple tests for the recommended length of time, that is when you can consider your overclock "stable". I, and many others, recommend a combination of tests.
Personally, I like 8 hours running Prime95 custom Blend mode. Choose Blend mode, then click on Custom. Input a min FFT of 512 and a max FFT of 4096. Enter an amount in the memory field that is equal to approximately 25% less than whatever your system is currently showing as free/not in use. Many supposedly stable overclocks will error out in this test showing it is not even reasonably stable. It is however possible for memory configuration or fault errors to cause this test to fail, highly likely in fact if something is wrong, but this should not be an issue if you are using the proper procedures because your CPU overclock should be established and verified as stable long before you ever do any overclocking of your memory or even setting it to its XMP profile settings. Default memory settings should always be used when first verifying the stability of your CPU overclock. If you prefer though, you can simply run the default Blend mode test in Prime version 26.6 or a newer version with an altered "local.txt" file that inhibits the use of AVX instructions being run since newer versions do not run realistic levels of AVX usage.
Next, if you can pass that, run the Realbench stress test for 8 hours. If you are only initially performing checkpoint testing along the way to a higher overclock, then you can probably get away with only 4 hours testing here but you may find later that at a higher clock you fail during an 8 hour test and may have to come back and run the full 8 hour test again at the lower clock speed to verify if the problem was actually only at the higher clock speed or if it did not in fact pass 8 hours at the lower clock speed/different voltage setting, as well.
If you can do those two things, and if you like, also run the recommended 12 hour Aida64 test, then I'd agree that for all intents and purposes you have a stable overclock BUT I will not agree that you can do any of this with that CPU, at that clock frequency, with THAT voltage setting. Likely, that voltage sensor value is what you are seeing within your monitoring utility, not what the ACTUAL setting in the bios is set at.
If you can show me the actual core voltage, LLC, offset and frequency settings in a bios screenshot, plus the results of these two stress tests I've outlined along with an HWinfo screenshot while running it, we can have a conversation. Otherwise, it's all just blah, blah, blah. Don't take this personally. It's what I expect of myself and everyone else as well. I would tell anybody, even somebody with a lot more experience than I have, something along the same lines because it's what I've been told in the past by overclockers with far more experience than you or I will likely ever have.
Facebook
Twitter
Instagram