The process of boosting CPU performance causes some confusion.
The process of boosting CPU performance causes some confusion.
82C seems quite elevated for a 7700K running at normal speeds with a 280mm AIO, though it isn’t extreme when using Cinebench. Ensure the fans on the radiator have enough space to move air (especially if set as exhausts), and consider boosting the pump speed to check its impact. Don’t stress too much about the mount unless the engine temps climb to 60-70°C or higher.
If the maximum temperature is reached, it’s unlikely any of these outcomes caused CPU throttling—possibly result #3, but none in the 80s. Performance won’t drop from the CPU until throttling hits around 100°C. The data shows a modest 7.x% boost and a 16% speed increase, which feels a bit weak (my 8700k usually gains ~15% with a 16% frequency rise). It probably isn’t thermal throttling on the CPU side if those temps are accurate. Could your VRM be limiting performance? Are you using a motherboard with cooling for the VRM parts? What’s the VRM temperature?
Intel processors will function at maximum output until the TJmax threshold, which is 100c. They typically don’t reduce performance until they reach that limit. Your highest score was 2461 at 4.9ghz, while your lowest was 2285 at 4.2ghz. This represents a 7.7% variation in score and a 16.6% difference in clock speed. The VRM refers to the power delivery system of the motherboard. Certain parts on the board produce heat, and if they become overloaded, they may reduce performance to maintain stability. #2 and #3 are nearly identical, differing by only 50mV and 9c. It’s likely that #3 underperformed due to thermal throttling during sensor updates. #1, #2, and #5 suggest the VRM is limiting power at higher voltages, not the CPU itself. In those cases, the CPU should operate within TJmax and shouldn’t throttle. Yet #5 matched #1’s performance despite being 14% faster, which is unusual. The difference lies in voltage: #1 was only 1.0v lower than #5, whereas #5 was 1.25v higher. Also, #4 outperformed #5 even without thermal issues or slower clocks. This inconsistency hints that your CPU isn’t throttling due to heat (except possibly #3). Assuming your measurements are reliable, consider running multiple tests with consistent settings—using tools like Cinebench R20 can help average results. For certainty, download XTU and perform a stress test to verify performance limits. Even if temperatures stay stable, you should confirm whether throttling is occurring.
Thanks for your feedback. I turned off XMP in the BIOS, which seems to have boosted my RAM performance despite the motherboard's limits. I experimented with different settings and got these updated stats: 1) 4.2 at 1000V, 2128°C max; 2) 4.8 at 1250V, 2402°C max; 3) 4.9 at 1275V, 2458°C max. No throttling or errors reported. XTU doesn't show any issues and performed well in tests.
Adjust your RAM speed appropriately, enable the XMP profile, and set the DRAM voltage to 1.35V manually. Use tools like A64 or Realbench to verify stability, monitoring temperatures and voltages closely. Also, look at what others are achieving at similar clock speeds and compare outcomes.