My 8700k Overclock results - 4.9ghz/1.24v - room for improvement?
My 8700k Overclock results - 4.9ghz/1.24v - room for improvement?
Prime 95 performs error verification for the CPU and memory. OCCT examines CPU and GPU as well as power supply for issues (voltage changes). Furmark, MSI Kombuster identifies GPU problems.
I'm currently running a real benchmark, and halfway through it, the bottom reports "instability detected" with "result hash match" in the left window. I attempted to click on showing results but got an error. I need to complete a full run at least once. Should I finish to see the outcomes or cancel now that Satana instability was detected?
You have the option to cancel whenever you notice mistakes, issues, or malfunctions.
I stopped the test and increased the voltage from 1.24 to 1.28. The temperatures reached 88 during the 15-minute thermal test. Is this acceptable? I didn’t proceed to the real bench because I wasn’t confident.
I lowered it slightly back to 1.25 and reduced the frequency to 4.8. Temperatures peaked at 82. I’ve successfully passed the real bench for two hours now. I’d like to return to 4.9, but the temperatures are still a concern. Although it might only reach 88 under maximum load, that seems unlikely in regular use.
Yes, it is too high. It is strongly advised to halt and reset your BIOS to the standard configurations, verify if a newer BIOS version exists and install it if available, then restart and gradually increase your overclock until you reach a stable point where thermal limits are not exceeded (no more than 80°C). It remains safe up to 85°C, provided you do not expose the CPU to immediate damage; however, continuous use at or near 85°C will eventually cause harm. If you cannot maintain below 80°C, your frequency is excessive, voltage too high, or cooling insufficient for the overclock. Keep stability intact with no errors in Realbench or other stress tests. A 4.9Ghz overclock would be considered extreme, as it exceeds the Intel base turbo by about 600MHz (roughly 1.2GHz above base and 200MHz above single core max), and even more so compared to other Coffee Lake CPUs. As a 6/12 processor rather than a 4/8 like previous generations, you won’t achieve the same gains this time due to the additional two cores (four extra logical cores). You’ll notice significantly faster heating when increasing by 500MHz on a 6/12 CPU compared to a 4/8 CPU. I followed the guide I previously shared and will proceed step by step. Rushing is not recommended, as it risks long-term damage to your processor or motherboard.
The situation is quite serious. The temperatures are exceeding safe limits, and the system is showing instability. It seems the overclock is too aggressive for the hardware, possibly due to insufficient cooling or voltage constraints. Adjusting frequency and voltage might help, but you should avoid pushing further until stability returns. Check your cooling setup and consider lowering the frequency again if needed.
If there are issues, they usually point to a few specific problems in a system that was working normally before the overclock. This is why I always suggest testing these on the original setup first, so you can confirm if the stability existed before changes. If the system was stable prior, it often suggests the frequency is too high, the sample CPU isn’t compatible, power delivery is insufficient (weak VRMs), or the voltage is too low.
For your case, you’re checking the voltage and thinking it’s too low. Increasing it usually raises heat, which might force you to lower the overclock. This is why newcomers sometimes assume they have a better sample than experienced overclockers, since they skip proper stability and thermal checks. If they followed the right process, they’d find they need more voltage for stability and often end up exceeding thermal limits, requiring them to adjust the multiplier until both thermal and stability requirements are met. 4.8Ghz could be achievable or not—try a slight voltage increase and retest.
If there are issues, they usually point to just a few things on a system that was working fine before the overclock. This helps confirm whether the problem started because of the overclock settings or the hardware itself. If the system was stable prior to the change, it often suggests the frequency is too high, the sample CPU isn’t compatible, power delivery is insufficient (weak VRMs), or the voltage is too low.
For your setup, I notice the voltage you have seems low. Increasing it could raise heat and force you to lower the overclock, which beginners often overlook when they assume their samples are better than others. Proper stability and thermal checks are essential before pushing higher.
4.8Ghz might be out of reach for you. You could try a slight voltage increase and retest, repeating until you find a balance between performance and stability.
Regarding thermal limits, going up to 85-90°C during stress tests is unusual. The chips usually throttle automatically to avoid damage. If you’re running gaming tests outside those conditions and keeping temperatures under 70°C, it might be acceptable. Deb8auer mentions up to 95°C in his guide, but he notes this could affect stability.
The majority of users will never encounter the temperatures you see during stress tests. Most software isn’t designed to keep the processor at full capacity. However, some are. As temperature rises, the likelihood of an error increases, along with the power required to maintain stability. Error chances grow with both speed and heat. Running at standard speed 95c is close to the maximum temperature, yet the risk of an error remains extremely low. Intel conducts thorough testing and categorization to ensure this. Once you begin overclocking, all those safeguards disappear, and you must adjust it yourself. Different chips handle overclocking differently—some excel, others struggle—it’s all about finding the right balance in the hardware. Or simply stick with stock performance.