I just bought a new Intel i5 8400, an MSI B360 Gaming Plus motherboard and 8GB DDR4 2400 RAM. I am using an Cooler Master Hyper 212 Evo.
I am wondering if the temperatures of the processor are good. I got about 61 degrees C with intel burn test, with stress level set to High. The room temperature was about 21 degrees C.
I used Arctic Silver MX-4 thermal paste.
Are those temperatures (61 degrees C maximum) normal if I use an aftermarket cooler, such as 212 Evo ?
Hello, mike69. Thank you very much for sharing your issue with the Intel Communities Team. I will be more than glad to assist you.
Those temperatures are actually very good during a stress test. Please keep in mind that the Tjunction (maximum temperature allowed at the processor die) of this processor is 100C so this processor is not even close to overheating. You can double check that information in this link: https://ark.intel.com/products/126687/Intel-Core-i5-8400-Processor-9M-Cache-up-to-4_00-GHz https://ark.intel.com/products/126687/Intel-Core-i5-8400-Processor-9M-Cache-up-to-4_00-GHz
I would like to add as well that the TDP (average power, in watts, the processor dissipates when operating at Base Frequency with all cores active under an Intel-defined, high-complexity workload) of this processor is 65W and, the http://www.coolermaster.com/cooling/cpu-air-cooler/hyper-212-evo/ Cooler Master Hyper 212 EVO will dispel 150W so you should not have any problem with the heatsink.
You should not worry about that 61C, as previously mentioned, those temperatures are fine and you should not have problems with the hardware either.
I hope this answers all your questions.
Hello, mike69. Thank you for your response.
Please verify if the processor is not damaged by removing the heatsink if possible. If the computer boots normally it should not be a problem since the leveler will push the Motherboard as well so the processor is not taking all the damage.
Thanks for the reply.
While I run Intel Burn test, the frequency of all cores is about 3800MHz, but when I do not run Intel Burn test the frequency is about 3900MHz.
Why the frequency is 3900MHz when I do not run the Intel Burn test ? It is normal to be about 3900MHz (please have a look at the attached image) or it should be 3800MHz ?
Hello, mike69. Thank you very much for your reply.
Please keep in mind that the processor will increase or decrease the frequency as it requires to. It is normal if the frequency varies even during a stress test. In this case, it seems that the Intel® Burn Test will only make the processor reach 3800MHz.
You can double check the health of your processor by running the Intel® Processor Diagnostic Tool. Here is the download link: https://downloadcenter.intel.com/download/19792/Intel-Processor-Diagnostic-Tool https://downloadcenter.intel.com/download/19792/Intel-Processor-Diagnostic-Tool
Thanks for the reply.
Can you please explain what is the meaning of "CPU Fan1 step up time" and "CPU Fan1 step down time" ?
I think that "CPU Fan1 step up time" is the time during the fan speed increases and "CPU Fan1 step down time" is the time during the fan speed decreases.
Please correct me if I am wrong. I also attached a screenshot to show you the settings in the BIOS.
I decided to use 1s for "CPU Fan1 step down time" to let the fan pull out all the hot air from the PC case, while I am in a gaming session.
Are those settings correct ?
These are rate parameters. They specify how fast the fan should respond to increases in temperature and decreases in temperature.
The values you have chosen seem pretty reasonable, but there are consequences to be understood. First of all, a very rapid response to temperature change can help ensure that temperatures do not get too high, but this can have a negative impact on psychoacoustics. This is obvious things - like fan oscillation, which gets annoying quickly - but this can be more subtle - like the tenet that, if it is changing pitch, it becomes noticeable and, if it is (too) noticeable, it can become annoying.
The processor fan is typically more acoustically noticeable than are chassis fans. Now, how fast the processor fan actually needs to spin to extract the heat being produced is dependent upon the temperature of the air that is being used. If the air is hotter, then more airflow is necessary to extract the same amount of heat. So, as the processor remains longer in a busy state (and the heatsink is being saturated with heat), the more important it is to lower the ambient air temperature within the chassis. This same rule also applies in the case of add-in graphics solutions. So, as CPU and GPU temperatures rise, not only do you want their dedicated fans to work harder, you want the chassis fans to work harder as well. Because the acoustic cost of using the chassis fans is lower and their use limits the (higher) acoustic cost of using the dedicated fans, there is an acoustic advantage to increasing chassis fan speed based upon CPU and GPU temperatures (in addition to the cooling requirements of other parts of the system).
Hope this helps,
Well, it would be difficult for me to answer this question with anything other than generalities. Every system is different (efficiency of thermal dissipation (including how well the TIM was applied, etc.), airflow patterns (including components added, locations of fans, affects of cables, etc.), temperatures of other components, etc.) and the (BIOS-level) capabilities offered for fan speed control configuration can vary (sometimes even when the underlying hardware capability is identical). But let's list some of these generalities and see where that gets you...
- Some PWM fans implement a minimum operational speed and will not go below it (unless turning off the fan). In my experience, this typically coincides with a minimum duty cycle of ~25%. Other fans, if you configure below this point, will try to implement the setting, but eventually you will reach a level where the fan will stall (I have seen few that went much below ~15% without stalling). At the same time, running the fan at (or above) some reasonable minimum level all the times is a good practice as it ensures a minimum level of dissipation. This allows more initial headroom when the processor is busy before the heatsink reaches saturation levels. I thus recommend that you implement a minimum duty cycle at 25%.
- As I mentioned before, the rate at which the duty cycle (and resulting fan speed) change can make fan noise more noticeable and, if it is more noticeable, there is a higher chance of this becoming a source of annoyance. [Aside: Often, leaving the fan at a much higher level and changing it rarely, while noticeable initially, will eventually be ignored completely (you simply become less and less aware of it). This isn't going to be a good practice if you are regularly listening to content, however.] What this means is that you want to minimize the rate of fan response increase/decrease so that it remains less noticeable. At the same time, however, you do need to address significant temperature levels. As temperatures rise, you can address this rise immediately or you can delay the response until absolutely necessary. The advantage of the former is that, if temperatures reach concerning levels, you are closer to the necessary fan response and there is less chance of reaching critical levels. The advantage of the latter is that you do not create a negative acoustic situation until you absolutely have to and, if the workload lessens and the temperatures drop before reaching these levels, then you have avoided the negative acoustics completely (or at least significantly). Obviously, there is a balancing point between these two schools of thought. My recommendation is a fan speed curve that looks roughly like this:
- My recommendation is to base the selection of the Tmin and Tmax values on your processor's Tcontrol temperature (for more information regarding Tcontrol, read my summary here: ). I recommend using Tcontrol as your Tmax value and Tcontrol-20 as your Tmin value. As my summary indicated, Tcontrol is typically in the low- to mid-80's. If you would like to shift the range a little bit higher and reduce acoustics, you could use, say, TJmax-10 (TJmax is the Maximum Junction Temperature). I wouldn't recommend going much higher than that, however.
- As for the Step-Up and Step-Down times, I think your values are a reasonable starting point. I would suggest running some tests, first using the Intel Processor Diagnostic Test as a baseline and then some heavier workloads - using something like Prime95 - to see whether any oscillation is occurring (especially at higher temperatures).
I hope that's enough to get you started...