Settings to stop CPU throttling automatically at certain temperature

[r3drocksr3d]

I have installed a CPU and played around with bios and windows settings but am unsure if the CPU is working or behaving as it should, so i have reset bios to default values and set windows power settings back to default values.

The CPU is now throtlling from ~4000ghz to ~1600ghz automatically frequently when the pc is idling. Under full load rendering a video it reaches its peak ~4000ghz ok at max temp 58c but will throttle sometimes to ~1600ghz, probably once every 5 minutes, and temperature drops as a result.

is this normal?

Is it possible to stop the pc from regulating the temperature automatically so it does not throttle under full load at 58c? The max working temperature for the CPU on the AMD website is 61c so it is not maxing out the temperature!

The cpu is amd fx8350.
I have Windows 10 installed.
I am monitoring the CPU core loads and temperatures with HWmonitor.

thanks for any help!

 

[kanewolf]

I would start by double checking your temps with AMD Overdrive. Remember that Overdrive shows the "thermal headroom" It will start at 40C and go DOWN. You were within a couple degrees, so you still could be thermal limited. I would also say you need a better cooler.

 

[TJ Hooker]

What motherboard and PSU?

I've heard lots of stories about the motherboard VRMs overheating on cheaper boards with 125W AMD chips, causing the frequency to be reduced even though the CPU temp is ok. Sounds like that might be happening to you.

As an aside, I've heard many utilities don't measure CPU temp properly on older AMD CPUs, the most reliable way is to measure thermal margin using AMD overdrive.

Edit: Also, what CPU cooler?

 

[kanewolf]

What motherboard and PSU?

I've heard lots of stories about the motherboard VRMs overheating on cheaper boards with 125W AMD chips, causing the frequency to be reduced even though the CPU temp is ok. Sounds like that might be happening to you.

As an aside, I've heard many utilities don't measure CPU temp properly on older AMD CPUs, the most reliable way is to measure thermal margin using AMD overdrive.

VRM overheating is especially a problem with AIO coolers because they don't have the CPU fan to move air over the VRMs.

 

[r3drocksr3d]

I would start by double checking your temps with AMD Overdrive. Remember that Overdrive shows the "thermal headroom" It will start at 40C and go DOWN. You were within a couple degrees, so you still could be thermal limited. I would also say you need a better cooler.

hi, thanks for the answer. I can see the core temps for the cpu, is this what i should take notice of? i could not see the general cpu temp in amd overdrive. they are still high, and the CPU throttles when all cores are on 100% load and at temps are about 57-59c.

If i manage to bring the CPU temp down with a better cooler do you think that will stop the throttling? If so can you recommend a cooler that will be an improvement on the stock cooler i am using, at a reasonable price?

I have had an after thought about testing the cpu with 50% or 75% load so the temp rises to maybe 50c approx to see if the cpu still throttles at a temperature not close to the limit, but do not know how to do that if it is possible!

Before i reset the bios and windows power settings to default, i had changed some settings so that the thottling was not happening while the pc was idle with no load on the cpu and the temp was low! But even with all the settings i changed i could not stop throttling at full load but the temp i measured was lower at about 57c full load!

In bios ‘cool and quiet’ was disabled, cpu fan speed low limit 500rpm, cpu fan set to ‘turbo’ and ‘core c6 state’ disabled.

Do you know what ‘cpu q-fan function’ setting does in bios?

thanks for everyone’s help and advice!

 

[r3drocksr3d]

What motherboard and PSU?

I've heard lots of stories about the motherboard VRMs overheating on cheaper boards with 125W AMD chips, causing the frequency to be reduced even though the CPU temp is ok. Sounds like that might be happening to you.

As an aside, I've heard many utilities don't measure CPU temp properly on older AMD CPUs, the most reliable way is to measure thermal margin using AMD overdrive.

Edit: Also, what CPU cooler?

Hi,

it is the stock cooler, i got the CPU off amazon, there is a lot of posts saying it is not good! Can you recommend a cooler?

thanks for any help!

 

[r3drocksr3d]

What motherboard and PSU?

I've heard lots of stories about the motherboard VRMs overheating on cheaper boards with 125W AMD chips, causing the frequency to be reduced even though the CPU temp is ok. Sounds like that might be happening to you.

As an aside, I've heard many utilities don't measure CPU temp properly on older AMD CPUs, the most reliable way is to measure thermal margin using AMD overdrive.

VRM overheating is especially a problem with AIO coolers because they don't have the CPU fan to move air over the VRMs.

Hi what does the thermal margin mean, is it the average temp over time?

The motherboard is a ASUS M5A78L-M/USB3
The PSU is 500W

The mothereboard temp can go up to 60c, would that be measuring the VRM’s

thanks in advance for all the info and help!

 

[kanewolf]

What motherboard and PSU?

I've heard lots of stories about the motherboard VRMs overheating on cheaper boards with 125W AMD chips, causing the frequency to be reduced even though the CPU temp is ok. Sounds like that might be happening to you.

As an aside, I've heard many utilities don't measure CPU temp properly on older AMD CPUs, the most reliable way is to measure thermal margin using AMD overdrive.

VRM overheating is especially a problem with AIO coolers because they don't have the CPU fan to move air over the VRMs.

Hi what does the thermal margin mean, is it the average temp over time?

The motherboard is a ASUS M5A78L-M/USB3
The PSU is 500W

The mothereboard temp can go up to 60c, would that be measuring the VRM’s

thanks in advance for all the info and help!

Thermal margin is the amount the CPU can heat up before bad things happen, like thermal throttling. I like to see 10C under max load. That ensures there is plenty of margin.

 

[TJ Hooker]

The motherboard is a ASUS M5A78L-M/USB3
The PSU is 500W

The mothereboard temp can go up to 60c, would that be measuring the VRM’s

thanks in advance for all the info and help!

The motherboard temperature listed is likely the chipset, you probably don't have any way of measuring the VRM temperature short of getting an IR thermometer. Once you verify that your thermal margin is ok (as described by @kanewolf), we will be able to conclude it is likely the VRMs. That is a low end board with poor VRMs, so that wouldn't be surprising. If that's the case, you could try adding VRM heatsinks (e.g. https://www.amazon.com/Enzotech-MOS-C10-Forged-Copper-Heatsinks/dp/B004CL89D8) and/or try to increase airflow over the VRMs.