Reduced clock speed to 3.7GHz on your 8700K cut temperatures to 40°C—Is that the highest it reaches?
Reduced clock speed to 3.7GHz on your 8700K cut temperatures to 40°C—Is that the highest it reaches?
Cooler: H115i RGB pro XT RAM 32GB 3200Mhz CL16 Board: ASUS Maximus Hero X Wi-Fi GPU:6800 XT CPU isn't listed. I planned to check it after my processor hit 99°C at 5.0GHz with 1.37v. When the system ran smoothly I was seeing about 78°C during load in games. Recently it began to reach 99°C and forced all fans for 10 seconds every two minutes while playing. Temperatures stayed above 85°C in games. CPU usage was around 80% during gameplay. I tried a 4.7Ghz at 1.35v and temperatures remained near 90°C. I downclocked the CPU to 3.7Ghz at 1.15v in hopes of dropping temps below 40°C. You can view the results below 40°C during full load and idle here: https://imgur.com/a/z7L3Wui The cooler pump is functioning properly; I can feel and hear it activating. It was swapped with Corsair three months ago and I was experiencing under 80°C during load for two months at 5Ghz. CPU fans are directly linked to the board; not the cooler. I ran them at full capacity for this test to ensure optimal cooling. You can see the pump operating at maximum speed. I experimented with different coolant pastes. The room temperature is 25°C. Are these readings indicative of what you'd expect at a base speed of 3.7Ghz? What causes the sudden 99°C spikes at 5.0GHz?
The water block fits securely on the motherboard, but there could be a contact problem. Or you might be applying too many volts. Check if the BIOS sets it to 1.37V and if your monitor shows that value for VCore in monitoring tools. VRVOUT is what you’re seeing reported. Your 8700k ran at 5GHz with about 1.30 volts and stayed cool—good performance for a solid chip. At 1.37V it’s pushing limits, especially under stress. What software are you testing? That affects stability checks. Your 9900k handled 5GHz at a slightly lower voltage and stayed manageable. At 1.35V it would overheat quickly, causing thermal throttling and performance drops. For a full custom build, aim for 5GHz at around 1.35V to avoid issues.
Underclocking a CPU using a custom liquid cooling setup seems wasteful. I concur, probably due to a connection problem.
Report a problem with your intel toothpaste. It seems to be faulty, but after using it for three years and applying LM, it still works well. Performance is 5GHz at 1.38v under daily use.
I began the evaluation with PRIME95 today. The thermal paste was swapped beforehand. The water block fit perfectly. Measurements were taken at 5GHz with a voltage of 1.32v. I reduced the voltage after your message. The device runs at 38°C. The Vcore value is 1.312V in HWINFO64. I plan to run the tests at normal settings once completed. I haven’t confirmed this yet since temperatures are high. I checked the "small FFTs" setting and stopped the test when I saw 99°C results: https://imgur.com/a/BzBl53u#8flVwL8 This was another "custom" run with AVX turned off: https://imgur.com/a/SeoJ8Si#rWbmdKF
Prime95 functions more like a performance tool rather than a stability tester anymore. For accurate results, refer to Asus real bench: https://rog.asus.com/rog-pro/realbench-v2-leaderboard/. Run the stress test, set RAM to the value shown (e.g., 8GB for 16GB), and ensure it completes all 8 hours without issues. Avoid exceeding 99c to maintain stability. Adjust the overclocking level if it causes excessive voltage or temperature spikes.