After around 10 minutes of playing demanding games, my frame rate drops for a few seconds roughly once every minute. MSI Afterburner shows my CPU dropping from its usual 3900–4600 MHz down to about 800 MHz during these stutters. My fans become very loud shortly beforehand, then briefly stop. The GPU usage also falls to 0%, presumably because the CPU is no longer supplying it with work.
My system has an RTX 3060, an Intel Core i5-11600K, a B560MX-E PRO motherboard, and 16 GB of RAM. Memory usage is usually around 14–15 GB while gaming. I lowered the BIOS power limit from 150 W to 90 W, but the problem continued. Before lowering the limit, CPU temperatures reached about 80–90°C; during the throttling episodes afterward, they were around 60–70°C. What could be causing the periodic 800 MHz throttling?
3 Answers
The first thing to verify is whether the CPU is hitting a thermal or power-protection limit. The loud fans before each stutter suggest the system is reacting to a temperature or motherboard sensor event. Monitor CPU package temperature, clock speed, and throttling flags with a tool such as Core Temp or HWiNFO while reproducing the problem. Also clean dust from the case and heatsink, check that the CPU cooler is mounted properly, and consider replacing the thermal paste if it is old.
Reducing the power limit may not address the trigger. Try temporarily lowering the CPU multiplier or maximum clock in Intel XTU and see whether the stutters stop. A monitoring utility can also show whether the cause is thermal throttling, power-limit throttling, or BD PROCHOT, which can force the CPU to reduce its speed because another component reports a problem. Change one setting at a time and keep a log while gaming for at least 15 minutes.
The useful next step is to capture the exact reason for the slowdown rather than relying only on temperature. Log CPU frequency, package power, temperatures, voltage, and throttling indicators during one of the episodes. If temperatures stay at 60–70°C while the CPU falls to 800 MHz, investigate motherboard power delivery, BIOS settings, the cooler and fan behavior, and external protection signals instead of assuming the processor is overheating.
Before lowering the power limit, temperatures were around 80–90°C. After lowering it, the CPU is only at about 60–70°C when the clock drops, so I’m not sure a normal temperature limit is the cause.