Software fails to detect overclocking capabilities.
Software fails to detect overclocking capabilities.
Hi everyone, I own an MSI PRO Z690-A DDR5 board with an i5-12600K processor. I've adjusted the BIOS settings to boost the P-cores to 5.2ghz and the E-cores to 4.2ghz. Using CPUID and HWMonitor, I notice the frequencies can spike from 801MHz up to 5210MHz (CPU state between 5% and 100%). That sounds impressive. The issue arises during stress tests—both HeavyLoad and CPU-Z show the CPU running below its advertised Turboboost numbers (4.9GHz for P-cores, 3.6GHz for E-cores). Why don’t stress tools use the overclocked settings I’ve applied? Do you know anyone who’s experienced this? Thanks ahead of time!
To check what’s causing throttling, examine the log output for specific error messages or patterns related to bandwidth limits. The reported value of 49.00 x 100.0 MHz suggests you may need to adjust the threshold in your configuration. If you’re unsure, compare it with expected values and look for anomalies. Your report file is a good reference—just change its extension to .txt and open it for more details.
Discovered these details about ring limits. Frequency is restricted by package-level PL2, PsysPL2 or PL3 power caps, and the ring's maximum voltage. Electrical design or other factors may also cause the interconnect frequency to be limited. Current consumption is capped by IA core frequency due to thermal effects.
I adjusted the OC settings back to 100 and enabled all cores instead of offset. The cores are now locked at their respective OC levels—5.2 for P-cores and 4.2 for E-cores—but they don’t lower below those values even when idle. When the system isn’t busy, the frequencies stay unchanged despite OS power settings ranging from 5% to 100%. The stress test is now using these OC frequencies, but temperatures are much higher than before. Idling runs at around 52°C instead of the previous 30°C, and a 2-second stress test pushes it up to about 95°C. I suspect the issue stemmed from setting the cores to run at the OC frequency rather than manually locking them at a fixed speed. That way, the high frequency causes overheating problems.
I apologize for the numerous responses regarding this issue. I believe I've resolved the problem. The original setup used turboboost offsets, but the auto offset was adjusting the cores to 52, 51, 51, 50, and 49 respectively. During testing, it consistently hit the lowest core without exceeding it. Now I manually adjusted each P-core to 52 and each E-core to 41 instead of relying on BIOS offsets. This change ensures the actual frequencies are used during stress tests. When idle, it maintains 30°C and 74°C under heavy load, and it drops to 798MHz then jumps to 5.2GHz when active. Temperatures remain within acceptable ranges.