A8 6600K over 100 C while rendering

[stancee]

Hey guys!
So, recently I built a PC by myself. Everything works fine on it, no problems detected.
The only problem is my cpu temp while rendering in 1080p (even on lower resolutions). It stays around 100C+.
*I am using the stock cooler, could that be the problem?
I have checked the heatsink to see if something is wrong with it, but everything is ok.
And when I am rendering, the cpu stays at 100% in task manager.

Specs:
CPU: AMD A8 6600K 3.9 ghz
RAM: 8GB Corsair (1600mhz)
Motherboard: Gigabyte F2A88XM-DS2
GPU: Integrated(8650D), stock frequency
HDD : WD Blue, 1TB

Here is a video while rendering in media encoder:

https://www.youtube.com/watch?v=EHk_taQXJYE&feature=youtu.be

As you can see, I hit 110C, and I am pretty sure that at this temperature the cpu could die in a few weeks, maybe months.
I am planning to buy in a few days a Deepcool Iceedge Mini FS (is it good?), but till then, can you tell me what else can I do to reduce the temperature?

 

[Insert_Nickname]

*I am using the stock cooler, could that be the problem?

Perhaps. I had a similar problem that I eventually traced to a combination of poor airflow, specific airflow around the APU heatsink and which memory speed is used. See this thread:

http://forums.anandtech.com/showthread.php?t=2365426&highlight=small+warning

At the least you should ensure a high rate of airflow around the APU heatsink. If you can mount a 120/140mm fan in the case side, it would likely help a lot.

Oh, and welcome to the forums...

 

[CuriousMike]

I'm not looking at the video, but my sons A10-5800k build, HWMonitor could not correctly identify the temps.
I went thru 3 programs before finding the right one.

 

[coercitiv]

Maximum operating temperature for that CPU is 74C. You are definitely not reaching 110C.

Is your CPU voltage set to stock? It seems to me it should be significantly lower than 1.31V, probably under 1.2V

 

[Phynaz]

There is no way that CPU is getting to 110c. It would shut itself down first.

Your temps are being reported incorrectly. This is because AMD doesn't publish their sensor data, therefore reading the temperaure of an AMD CPU via software is all a guessing game.

 

[monstercameron]

yeah richland had errors with temp monitors...

 

[stancee]

Okay, let me get this right. When my pc is on idle the temp is around 40 degrees, which is normal. But when I start doing something, the sensors are starting to give false information?
And yes, my cpu voltage is on stock.

 

[sm625]

The 40 deg is probably also wrong. When you load it up and let it run for a few minutes, reach inside and touch the base of the heatsink. If its really 100C then that sucker is going to be smoking hot. But I bet you it is merely lukewarm. If you're feeling extra experimental, you can always use a thermometer in conjunction with a frying pan and a stove to approximate the temperature of your HSF under load. From personal experience, a 65C heatsink will be too hot to touch for more than 5 seconds. A 75C heatsink will be too hot to touch for more than 3 seconds.

 

[Phynaz]

Okay, let me get this right. When my pc is on idle the temp is around 40 degrees, which is normal. But when I start doing something, the sensors are starting to give false information?
And yes, my cpu voltage is on stock.

Read my post again. I never said the sensors are giving wrong data. I said AMD hasn't published the sensor data.

Big difference between what I wrote and what you interpreted I wrote.

Your CPU will not run at 110c. It would be destroyed.

 

[stancee]

sm625, you were right. The heatsink is warm, but not over 60c.
Thank you guys for all the information about this!

 

[tweakboy]

The CPU wont die,, it will just fry over time.

 

[zir_blazer]

Its either the Heatsink incorrectly installed, or incorrectly reported temperatures. For the latter you can try with CoreTemp, which is specialized on only doing temperature readings. Also, you should pay attention to performance or unexpected shutdowns. If it ain't throttling, nor shutdowns or stability issues, I wouldn't believe it is really overheating. And if the Heatsink was improperly installed, you need to unmount it, clear the Processor and the Heatsink base, reapply paste and install it again. A mid end Processor shouldn't need aftermarket cooling.

Airflow is important as components starts to clutter, but for a simple computer in an ATX cases with no discrete Video Card, you don't even need Case Fans. After several hot 30°C+ summers with several machines with no Case Fans I can say than low end and mid end parts doesn't really needs them.

Maximum operating temperature for that CPU is 74C. You are definitely not reaching 110C.

74°C should be TCaseMax. AMD's TCaseMax is usually wrongly quoted as the maximum operating temperature, but it is not. TCaseMax should be similar to Intel definition:

Tcase is the temperature measurement using a thermocouple embedded in the center of the heat spreader. This initial measurement is done at the factory. Post-manufacturing, Tcase is calibrated by the BIOS, through a reading delivered by a diode between and below the cores.

Tcase Max is the maximum temperature that the Tcase sensor should reach. Both Tcase and the thermal specification information can be found on the Intel web site.

From the times of Socket 939 Opterons that people wrongly thinks than on AMD TCaseMax = Maximum operating temperature. This was thanks to AMD own mistakes on some documentation (I recall AMD Compare wrongly stating so). It was fun seeing people panicking over low-leakage Opterons 1xx with a 53°C TCaseMax reaching 60°C or so and desperately trying to cool it.
The actual maximum operating temperature should have a definition similar to Intel TJunction, which is when throttling kicks in as an emergency measure and before automatic shutdown. And AMD doesn't reveals that value as far that I know.

The 40 deg is probably also wrong. When you load it up and let it run for a few minutes, reach inside and touch the base of the heatsink. If its really 100C then that sucker is going to be smoking hot. But I bet you it is merely lukewarm.

This does not work like that. That is actually a bad sign. If there is bad contact between the Heatsink and the Heatspreader, the Heatsink will NOT be warm as there isn't enough heat transferred. You will have a Processor overheating with a cold Heatsink.

Your CPU will not run at 110c. It would be destroyed.

Nope. Processors are much more heat tolerant than most people thinks. I had a faulty Athlon 64 Venice in which I confirmed that shut down temperatures are 120°C. I suspected than it could be than the TIM between the die and the Heatspreader (Yes, like Ivy Bridge and Haswells) was drying according to some Posts I read which people that had mysteriously overheating Athlons 64, due to the degradation over time which I saw with it. I delidded it but never managed to purchase an epoxy type TIM to put it back again, and soon after that computer was replaced.
I'm still curious if someone had similar issues with a 90nm A64 and if they managed to fix it or not. But seriously. If you don't hit shutdown temperatures, you don't really know how much punishment a Processor supports.

 

[Phynaz]

He's not running an A64.

 

[coercitiv]

This does not work like that. That is actually a bad sign. If there is bad contact between the Heatsink and the Heatspreader, the Heatsink will NOT be warm as there isn't enough heat transferred. You will have a Processor overheating with a cold Heatsink.

Nope, he would have a throttled CPU, which is not the case as shown in the video.

 

[coercitiv]

stancee, have you modified any BIOS settings related to CPU operation?

It seems to me your CPU is operating with the voltage value of it's highest turbo speed (1.35V) rather than the significantly lower value for base speed (likely under 1.2V). That wouldn't fully explain the temperature readings you're getting, but it would shine some light on the values for Total CPU Power in the video.

Could you also check the CPU temp indicated by the BIOS?

 

[stancee]

stancee, have you modified any BIOS settings related to CPU operation?

It seems to me your CPU is operating with the voltage value of it's highest turbo speed (1.35V) rather than the significantly lower value for base speed (likely under 1.2V). That wouldn't fully explain the temperature readings you're getting, but it would shine some light on the values for Total CPU Power in the video.

Could you also check the CPU temp indicated by the BIOS?

Well, I have once overclocked it to 4ghz and allowed the turbo mode. But the pc shut down itself when I was in a game, so I put the settings on default, with turbo mode off. It's on stock now.
I have checked the bios temp. It stays around 40-50c (I still had those temperatures before overclocking)

 

[coffeejunkee]

I would say the cpu 0 package temps are not read out correctly. I would focus on the cpu 0 temp which is fine at 60C. Seems about right to me for something intensive like rendering a video.