F5F Stay Refreshed Hardware Desktop Are high CPU temperatures affecting performance on another motherboard?

Are high CPU temperatures affecting performance on another motherboard?

Are high CPU temperatures affecting performance on another motherboard?

Pages (2): Previous 1 2
M
Mclovin564
Junior Member
34
04-11-2020, 05:49 PM
#11
5700g, operating at 3.8GHZ base clock; boost clocks commence above this point, delivering roughly 5.GHz across all cores and up to 4.6GHz when a single core operates under heavy load.
I still believe a Wraith cooler from Cooler Master is merely adequate for everyday use without forcing the CPU to reach full capacity, such as what Prime95 achieves. Unless the processor is severely weak, no game or most demanding applications will push it this far for extended periods. Most standard coolers are designed for typical performance rather than peak demands. The next generation, like 5800X, already lacks such a cooler despite sharing similar architecture but running at about 100MHz higher. This illustrates just how closely performance is constrained. I would prefer upgrading to something better—something available for around 30-50 dollars—that could truly unlock full potential, especially since Ryzen models are highly sensitive to heat. Overclocking isn’t necessary here; instead, lowering voltages and using CO (Curve Optimizer) in negative mode can help. Typically, this involves setting values between -10 and -30, allowing the system to adjust voltage dynamically based on boost needs. This should be accessible via BIOS options: Advanced > Overclocking > PBO. It’s not an overclocking process despite its appearance.
To assess the right configuration, monitor not just temperatures but also core and package temps alongside frequencies while running typical programs or benchmark tools like Cinebench or Ryzen Master. These provide basic insights.
OCCT is another excellent tool for testing, offering comprehensive data.
PS. It’s unclear which program you’re using to track temps, as results can vary—some tools average across cores, others highlight the hottest single core, potentially differing by 20°C or more. HWinfo offers the most detailed information, while CPUtemp provides the least.
M
Mclovin564
04-11-2020, 05:49 PM #11

5700g, operating at 3.8GHZ base clock; boost clocks commence above this point, delivering roughly 5.GHz across all cores and up to 4.6GHz when a single core operates under heavy load.
I still believe a Wraith cooler from Cooler Master is merely adequate for everyday use without forcing the CPU to reach full capacity, such as what Prime95 achieves. Unless the processor is severely weak, no game or most demanding applications will push it this far for extended periods. Most standard coolers are designed for typical performance rather than peak demands. The next generation, like 5800X, already lacks such a cooler despite sharing similar architecture but running at about 100MHz higher. This illustrates just how closely performance is constrained. I would prefer upgrading to something better—something available for around 30-50 dollars—that could truly unlock full potential, especially since Ryzen models are highly sensitive to heat. Overclocking isn’t necessary here; instead, lowering voltages and using CO (Curve Optimizer) in negative mode can help. Typically, this involves setting values between -10 and -30, allowing the system to adjust voltage dynamically based on boost needs. This should be accessible via BIOS options: Advanced > Overclocking > PBO. It’s not an overclocking process despite its appearance.
To assess the right configuration, monitor not just temperatures but also core and package temps alongside frequencies while running typical programs or benchmark tools like Cinebench or Ryzen Master. These provide basic insights.
OCCT is another excellent tool for testing, offering comprehensive data.
PS. It’s unclear which program you’re using to track temps, as results can vary—some tools average across cores, others highlight the hottest single core, potentially differing by 20°C or more. HWinfo offers the most detailed information, while CPUtemp provides the least.

F
fruitypie121
Member
64
04-11-2020, 07:34 PM
#12
I can clarify more about reducing voltages and applying the curve optimizer. You can find helpful guidance online, including tutorials and videos that walk through these steps properly. Checking reliable sources will ensure you use them effectively.
F
fruitypie121
04-11-2020, 07:34 PM #12

I can clarify more about reducing voltages and applying the curve optimizer. You can find helpful guidance online, including tutorials and videos that walk through these steps properly. Checking reliable sources will ensure you use them effectively.

F
freddarley12
Junior Member
48
04-12-2020, 12:35 AM
#13
There isn't much to it actually. You can start with all core for the beginning and then adjust to negative values around -15. Use all core benchmarks like CineBench 23, record the results and temperatures. If it passes without issues up to -20, that's good. I believe -30 is the limit for the 5000 series Ryzen, but you have a solid sample for that. You might want to check this or any from the 5000 series, even though it's closest to 5800(x). Anything that works for it will work for yours.
F
freddarley12
04-12-2020, 12:35 AM #13

There isn't much to it actually. You can start with all core for the beginning and then adjust to negative values around -15. Use all core benchmarks like CineBench 23, record the results and temperatures. If it passes without issues up to -20, that's good. I believe -30 is the limit for the 5000 series Ryzen, but you have a solid sample for that. You might want to check this or any from the 5000 series, even though it's closest to 5800(x). Anything that works for it will work for yours.

Pages (2): Previous 1 2