Check if the Fx 8350 with MSI 990FX gaming overclocking is compatible and beneficial.
Check if the Fx 8350 with MSI 990FX gaming overclocking is compatible and beneficial.
Read the link I sent you carefully; the basics remain unchanged even if the wording in your bios differs. You still have automatic voltage control, which likely explains the high temperatures despite the low overclock.
The only steps required are to turn off AMP, C1E, cool a quiet mode, and SVM. Set LLC to high or ultra.
Safe temperature limit is 62°C core/70°C socket. Use AOD to verify thermal margins.
Max voltage should not exceed 1.55V.
Adjust the multiplier by one notch at the start, then perform 20 minutes of P95 SmallFFTs to assess stability. If it fails, increase voltage by 0.01-0.03 steps and re-test. Continue until you hit thermal/voltage limits.
For a definitive check, run Prime95 SmallFFTs for at least 3 hours to evaluate stability. Some recommend longer durations (8, 12, 24+ hours), but be mindful—real-world performance may differ. Prime95 doesn't fully reflect actual CPU behavior.
Benjiwenji:
Follow these steps: turn off AMP, C1E, cool down, and SVM. Set LLC to high or ultra.
Safe maximum temperature is 62°C core / 70°C socket. Check thermal margins with AOD.
Limit maximum voltage to no more than 1.55v.
Add one bump multiplier at the start, then perform P95 SmallFFTs for 20 minutes to verify stability. If it fails, adjust voltage by small increments of .01-.03 and retest. Continue until you hit thermal/voltage limits.
For a final check, run Prime95 SmallFFTs for at least 3 hours to ensure stability. Some recommend longer durations (8, 12, 24+ hours), but be mindful that even stable OCs may not last forever.
Prime95 doesn't fully reflect real CPU behavior.
As for the results, I increased it to 4.3ghz and observed the following:
Reading from HwMonitor showed the package temperature capped at 32°C and the socket at 73°C.
With AMD overdrive, it reported a thermal margin of 37°C for the 32-core setup.
Current voltage is 1.408v; I'm unsure if it can go higher. Should I lower the voltage or continue increasing it as the temperatures approach the stated limits?
Appreciate the advice.
Using AOD CPU thermal margin indicates a 32°C buffer before reaching the thermal limit or throttling, which is considered overheating. Continue pushing the CPU until thermal margins drop to single digits during stress testing (ensure positive values, e.g., -3°C = 3°C overheating), or it reaches 1.55V, whichever happens first.
Use AOD CPU thermal margin. A 32c margin indicates up to 32c before it hits the thermal limit or throttles, which is overheating. Keep increasing the CPU until your AOD becomes negative on a positive scale (for example -3 equals 3c overheating). Or aim for reaching 1.55v, whichever happens first.
Avoid negative outcomes. Focus on single-digit values on the positive side. The thermal limit is zero. If the multiplier adjustment causes failure in the 20-minute P95 SmallFFTs, adjust voltage in small steps from 0.01 to 0.03 and recheck. If it fails again, increment voltage until success or reach the maximum of 1.55v. If successful, increase multiplier and repeat the process.
Don't proceed to negative values. Focus on single-digit positive numbers. The thermal limit is 0. When the multiplier fails the 20-minute P95 SmallFFTs, adjust the voltage in small steps between 0.01 and 0.03 and recheck. If it still fails, increase the voltage until it works or reach the 1.55V thermal limit. If successful, increase the multiplier and repeat.
Maximum thermal margin can be around 70°C? After adjustments. For starting at 1.4 volts on a 4.3GHz chip, should I lower it slightly and test before increasing?
Ensure thermal margin in AOD does not reach zero or become negative. Any amount above zero is acceptable. Avoid reducing voltage; 1.4v is acceptable unless overheating occurs. Continue increasing the multiplier until it fails the test.