FX-6300 experiencing changing CPU speed when under load, but stable during idle. Assistance needed?
FX-6300 experiencing changing CPU speed when under load, but stable during idle. Assistance needed?
with the same length and structure:
Hey Everyone!
I'm having trouble figuring out why my CPU frequency changes when I run stress tests.
Below is my current configuration:
AMD FX-6300 CPU
Cooler Master Hyper Evo 212 CPU Cooler
ASUS M5A78L-M/USB3 Motherboard
8GB DDR3 1600MHz
EVGA NVIDIA GTX 960 4GB VRAM
500W Corsair CX500 PSU
Samsung 850 EVO 250 GB SSD
I'm using Prime95 Blend for stress testing, CPU-Z to verify Core Speed, Multiplier, and Voltage, as well as benchmarking each profile. Finally, AMD OverDrive is tracking the temps.
I began overclocking from stock Core Speed 3.5Ghz and Core Voltage 1.25V by adjusting the core multiplier in BIOS until I failed Prime95 for 30 minutes. Then I increased the voltage slightly, then the multiplier further, and so on. Eventually, I reached stable performance at 4.6GHz with a voltage of 1.45V.
At that stage, I was seeing a minimum of 15.0C thermal margin (20.0C average) in AMD OverDrive.
I've read that a Thermal Margin of 0C means a core temperature of 70C when using AMD OverDrive, and the ideal max core temp for the 6300 is 60C, so I'm targeting a 15C margin, which keeps me around 55C. That's why I stopped once I hit 4.6GHz.
I've recorded temperatures, Prime95 errors, and CPU-Z benchmarks at each 0.1GHz increment.
In any case, I'm planning a 24-hour Prime95 test, but I wanted to clarify this beforehand.
My Core Speed, Multiplier, and Core Voltage fluctuate when stress testing, changing every few seconds, whereas they remain steady during idle.
Currently, I have AMD Turbo Core off and Cool N Quiet off with 22x Multiplier and 1.4V voltage, reference clock at 200MHz (4.4GHz), aiming for a minimum thermal margin of 18C (20C average) after a 2-hour stress test with no errors.
I've tested with both Cool N Quiet off and on, but it doesn't seem to affect idle or stress performance—possibly due to Windows High Performance Power Setting minimum processor speed at 100%.
I also tried adjusting FSB/Bus Speed/Reference Clock up to 205MHz, but kept it at stock 200MHz for now. It didn't influence the fluctuations.
The changes look like this... At idle, no shifts occur; Core Speed stays steady at 4.4GHz.
Once stress testing starts, within seconds—before AMD OverDrive shows temps—I see the multiplier jump from 22x to 15x and back every few seconds. Core Speed and voltage follow accordingly.
On one level, since it changes even with default settings (3.5 GHz, with or without turbo) during stress tests, I suspect this isn't a temperature throttling issue (also because my thermal margin consistently stays above 15C [30C at stock speeds]), so I should aim for around 55C Core Temp.
That's why I settled on a 15C margin.
I've also experienced black screens and reboots at 0% and 100% core speed, so I reverted to Auto.
I experimented with FSB/Bus Speed and Reference Clock settings (up to 205MHz), but decided to stick with the stock 200MHz for now. It didn't impact the behavior.
The pattern repeats like this...
- At idle: no fluctuations, stable Core Speed
- Under stress: multiplier shifts rapidly every few seconds, matching temperature changes
Regarding your questions:
1) Why does my Core Speed vary during stress tests?
2) Why does throttling happen at 4.5–4.6GHz despite a 15C margin? Could it be a motherboard issue or insufficient power delivery?
3) What adjustments can I make to reduce these fluctuations?
Looking forward to your feedback! I'm glad I joined the community. Let me know if you need more details.
Well, since no one had any solutions, I stayed up late attempting a fix. I installed Intel Burn Test and attempted stress tests using it. What I discovered is that (I BELIEVE) the stress tests themselves are responsible for the fluctuations during testing. When I run IBT on standard, the variations happen similarly to Prime95. On Maximum, I see a long stretch of steady 4.4GHz before it starts to change again. I also began monitoring the CPU speeds in the Windows Task Manager chart, observing consistent performance at higher stress levels. Additionally, I adjusted my fan settings to maintain lower temperatures, so I plan to try clocking...
Well, since no one offered any solutions, I spent the night working on it. I installed Intel Burn Test and attempted stress tests using that tool. What I discovered is that the fluctuations might actually be caused by the stress tests themselves. On standard settings, the variations happen similarly to Prime95. With Maximum settings, I see a long stretch of steady 4.4GHz before any changes occur again. I also began monitoring the CPU speeds in the Windows Task Manager and saw consistent performance at higher stress levels. I adjusted my fan settings to maintain lower temperatures, which could improve stability further later. For now, I have a stable system—based on Prime95’s 24-hour test and IBT’s 20-run standard—running at 200MHz Base Clock, 22x CPU Multiplier, and 1.40V CPU Voltage, with a minimum thermal margin of 18.0°C (core temps around 52°C). I’m hoping this information will be useful for others in the future.