Throttling observed in Athlon 860k during low temperatures when under heavy workloads
Throttling observed in Athlon 860k during low temperatures when under heavy workloads
I'm working hard to reach a steady 4.4Ghz from my CPU, but it's proving quite challenging. I'm convinced the issue lies with me, even though the 860k isn't considered a top-tier overclocker, 4.4Ghz should be within reach. I haven't come near overheating issues.
Right now, I manage to hit 4.3Ghz without any trouble. The Prime95 temperatures remain in the mid to high 50s after several hours, without any adjustments beyond changing the multiplier.
The real difficulty comes when I try to increase the last 100Mhz setting. That's where my understanding stops. If I adjust the voltage manually, unexpected behavior occurs. Lowering it causes instability, while raising it leads to higher voltages than expected. Additionally, the displayed voltages in OHWmonitor are consistently higher than intended—sometimes reaching up to 1.55+ when set to 1.5. The CPU also throttles despite temperatures staying in the low 50s. When I check all temperatures via openhardwaremonitor, nothing seems out of the ordinary.
I'm assuming that achieving 4.4Ghz is possible because my temperatures are so low, but I might be missing a necessary setting or misconfiguring something.
I suspect the VRM could be restricting performance, though I have one of the top-rated FM2 boards available, not the extremely costly Crossblade Ranger, which offers similar power with better cooling.
I'm aware it's likely in there, but I need guidance to locate it. Any suggestions would be appreciated.
I believe I understood it correctly. Found a useful reference in the googly. It seems APM was throttling it to maintain power usage below the TDP. Not entirely certain, but it appears stable and reduces throttling significantly. The clocks remain somewhat erratic in Prime95, fluctuating between 4400 and 3500 quite often, which is better than gradually dropping into the 2k range. CPU temperature stayed around 56°C as well.
I ran multiple benchmarks and achieved improved scores overall.
I plan to check if lowering the voltage helps, since it's still set at standard volts. OHWmonitor reports they can reach up to 1.56v. I might also test other monitor programs to verify accuracy.
For anyone interested, I...
I believe I understood it correctly. I found a useful reference in the googly. It looks like APM was limiting its performance to maintain power usage under the TDP limit. Not entirely certain, but it appears stable and throttling less now. The clocks remain somewhat erratic in Prime95, fluctuating between 4400 and 3500 quite often, which is better than gradually dropping into the 2k range. CPU temperature stayed around 56°C as well.
I ran multiple benchmarks and saw overall score improvements.
I plan to check if lowering the voltage helps, though it's still at standard voltage, although OHWmonitor reports it could rise up to 1.56v. I might also test other monitor programs to verify accuracy.
For those curious, I turned off APM, used C6 mode, and set LLC to extreme. Voltage remains at the stock 1.475 in offset mode (which seems excessive, but stock is stock...).