No temp change after overclock
No temp change after overclock
Hi, your FX 8350 was overclocked to 4.5ghz and stress tested. Both max temperatures stayed around 58°C with stable readings during the test at 4.0ghz. Temperatures gradually settled between 49-58°C. Checking for differences is recommended. Idle temps are 25-35°C, sometimes reaching 40°C before dropping quickly. CPU-Z and HWmonitor show 4.5ghz, and CPU vcore is 1.4375. Ratio set to 22.5. FX 8350 @4.5 with Gigabyte 970a-D3P and Corsair H60, 16GB at 1333mhz.
I was genuinely puzzled by how much the FX CPU limited your graphics card performance. When I opened the first link and saw Unigine, I assumed the graphics card would be the main issue. But I was taken aback by the noticeable difference. I considered using a Cinebench CPU benchmark instead.
Regarding voltage, it really depends on the specific CPU. A common guideline is that lower stock voltages at regular speeds tend to allow better overclocking, as less voltage is needed to reach higher speeds. This isn't an absolute rule, but it generally holds true. For the FX 8350, most people agree that 1.5V is the safe upper limit.
I've discovered that most temperature monitoring tools don't work well with AMD's FX CPU, particularly when the system is idle. The most effective method is to use Overdrive and check the Thermal Margin. This value indicates how close the CPU is to throttling. It functions inversely to absolute temperature—higher numbers mean cooler operation. When the Thermal Margin nears zero, the CPU is very close to throttling.
To confirm functionality, first ensure you weren't throttled during load in both cases. If not, compare performance under different settings using a benchmark. This helps identify any performance variations. Generally, synthetic tests are needed to clearly show temperature changes.
Temperatures remain relatively stable with clock speed changes; noticeable rises only appear when voltage is increased beyond the core frequency. If you didn't need to boost Vcore from 4.0 to 4.5GHz, you likely won't see significant temperature increases.
I’ve discovered that most temperature monitoring tools don’t work well with AMD’s FX CPU, particularly when the system is idle. The most accurate method is to use Overdrive and check the Thermal Margin. This value indicates how close the CPU is to throttling; a higher number means it’s cooler. It functions in the opposite way to absolute temperature—higher numbers suggest lower heat.
To confirm if the readings are reliable, first ensure you weren’t throttling during load tests under both conditions. If not, compare performance before and after overclocking using a benchmark. Often, synthetic tests are needed to clearly see any changes.
Temperatures only shift slightly with clock speed; noticeable rises happen mainly when voltage is increased. If you didn’t need to boost Vcore from 4.0 to 4.5GHz, you probably won’t see significant temperature changes.
Thanks for your prompt reply. I checked AMD Overdrive and my thermal margins seem stable (~35-45°C). After overclocking in Unigine Valley, I observed clear differences. Here are the links if you want to review:
https://gyazo.com/48af169c2e5764fc268b883a9b313033
https://gyazo.com/721481c7bbcc329698773d3e2652afca
https://gyazo.com/247573c258c16e9cf1dadd00bbb4baa2
I’m comfortable with the current speed and temperature, so would I consider pushing further? If yes, what voltage levels seem to work best for overclocking? Or should I stick with what I have?
your temperatures could stay relatively stable even if your fan speed settings are more aggressive.
(when I overclock, I keep my fans running at full speed continuously so I can gauge how much room temperature is available for increasing the clock. Cooling performance is usually the biggest constraint in this process, although it's not the only one) and in case the system crashes, it lets me cool the system down before I restart.)
I was genuinely puzzled that the FX CPU limited your graphics card's performance so much. When I opened the first link and noticed it was Unigine, I assumed the graphics card would be the limiting factor. But I was taken aback by the noticeable difference. I considered using a Cinebench CPU benchmark instead.
Regarding voltage, it really depends on the specific CPU. A common guideline is that lower stock voltages at normal speeds tend to perform better for overclocking, as it requires less voltage to reach higher frequencies. This isn't an absolute rule, but it often holds true. For the FX 8350, most people agree that 1.5V is the safe upper limit, with some going up to 1.55V. However, this safe voltage can vary based on temperature and cooling efficiency.
When attempting overclocking, avoid increasing voltage until you notice instability. Increase the multiplier until you lose stability, then adjust the Vcore until you regain control. It's important never to exceed the necessary voltage for your core clock frequency.
When reviewing your benchmarks once more, it seems there is more happening than initially thought. Usually, a higher percentage overclock leads to a smaller boost in benchmark scores. Especially here, since the benchmark is graphics-focused rather than CPU-based, increasing the GPU from 4.0GHz to 4.5GHz actually gives a 12.5% improvement. This contradicts my expectations. Typically, you only get close to a one-to-one gain, though that’s uncommon. Sometimes other system components can pull down the score from a perfect result. In your situation, I would have anticipated around a 10% rise in Unigine with a 16.66% overclock. Did you also overclock your graphics card?
The benchmarks seem to show a different trend than expected. Generally, a higher percentage overclock leads to a smaller percentage boost in scores, especially for graphics tests. In your scenario, going from 4.0GHz to 4.5GHz resulted in only a 12.5% increase, whereas Unigine improved from 2501 to 2917—a 16.66% gain. This is contrary to what I anticipate. Usually, improvements stay around a 1-to-1 ratio, though that’s uncommon. Sometimes other system components can lower the score from perfect performance. In your situation, I’d have anticipated about a 10% Unigine boost with a 16.66% overclock. Have you also overclocked your graphics card? Your GPU was increased from 1279mhz to 1350mhz during testing. How would you tell when instability appears so you can adjust the vcore accordingly? Would running a stress test help identify crashes or issues?
In all overclocking scenarios, any adjustment—whether changing core clock speed, voltage, or both—requires a stress test to ensure stability. For CPU overclocking, increasing the multiplier followed by booting the OS and running a stress test with tools like Prime 95 Small FFT will push the processor beyond typical loads. If it remains stable under these conditions, it should perform reliably in normal usage. Stress testing also reveals extreme temperature situations.
A comparable method applies to GPUs; after adjusting core clock, memory clock, or both, executing a video stress test such as Unigine will determine stability. Stable performance here suggests compatibility with gaming.
Instability manifests as erratic behavior. For CPUs, this might include BSODs, crashes, application freezes, application errors, or rounding issues in Prime 95. For GPUs, instability could involve driver crashes, game crashes, visual artifacts, rebooting, or persistent BSODs.
In all overclocking situations, whenever you adjust the core clock speed, voltage, or both, it's essential to follow up with a stress test to ensure stability. For CPU overclocking, you would increase the multiplier, then launch the OS and perform a stress test using tools like Prime 95 Small FFT. This will push the processor beyond typical usage, and if it remains stable under those conditions, it should be stable in regular use. Stress testing also provides insight into extreme temperature scenarios.
A comparable method applies to GPUs: after any change in core clock, memory clock, or both, run a video stress test such as Unigine. If the GPU handles it without issues, it should perform well in games.
Instability manifests as erratic behavior. For CPUs, this might include BSODs, crashes, application errors, or rounding issues in Prime 95. For GPUs, instability could involve driver crashes, game crashes, visual artifacts, rebooting, or BSODs.
Thanks for all the guidance and information! You've picked up a few valuable insights!