Folding Forum

Hey, so I just began running F@H on my home PC, and started with "Medium" as that's the recommended setting for most users. However, I found that I had CPU core temps actually hitting 100C on this setting, which seems abnormal.

My rig isn't the newest, but I am running an i7 4790k and a GTX 970 (temps only climbed to 64C before I dropped to light due to CPU temps). Cooling-wise, I have a Noctua U12S 120mm single fan tower on the CPU, two front intake 120mm 95cfm fans, 1 120mm 95cfm rear exhaust fan, 1 side 65cfm intake fan and two top 65cfm exhaust fans. The cooling components are relatively new (bought and installed within the last 6-7 months), and thermal paste was re-applied (a small pea sized drop) after appropriate cleaning with a microfiber cloth and 90% isopropyl within the last 6-7 months as well.

Normal idle CPU temps are mid to high 30s C, most games push temps at most into the low 60s C. F@H on light sits comfortably in the low to mid 60s as well. Is the temperature jump I'm seeing on medium typical, or do I need to check on ways to optimize airflow?

I am using Speccy to monitor temps, if that matters.

Hi Dan and welcome to the forum.

That seems quite unusual, particularly with that nicer cooler. I'd guess that there is dust in the radiator or something that is interfering with the cooling. Cleaning it should help.

If cleaning it doesn't help much, then I'd recommend setting up a custom setup that is more focused on the graphics card instead. Find the Folding@home icon in the taskbar down by the clock, right-click, and select Advanced Control. Once the FAHControl tool is open, move the power slider back to Full. Then click Configure in the upper-left corner, go under the Slots tab, and click on the CPU slot. At the top of the pop-up window, set the number of threads to 2, then hit Okay and then Save to update the configuration. This should leave your GPU fully utilized (and extremely productive for Folding@home) while lowering the usage on the CPU to keep the temperatures down. That CPU has 4 real cores and 8 hyperthreaded cores, so with 3 cores active, you should see less than 50% CPU utilization in Task Manager.

Since you recently changed your thermal solution I'd suggest limiting the cores F@H uses to avoid those temps.
Or maybe try messing with fan curves.

Welcome to Folding@Home!
Here is how to select more or less threads to execute:

In the taskbar to the lower right of the screen, you should see a F@H molecule icon, click it (you may need to click an Up Arrow to see it ^)

The second item in this menu is Advanced Control, click it

On this screen to the left is a Configure button, click it

Now you get a screen with a Slots tab, click it

On this white field should be a cpu item, click it and then click edit

By default F@H set the number of CPUs to -1 meaning let the software decide.
You can enter any number from 1 to the number of threads your CPU supports.
If you have GPUs, F@H reserves one CPU per GPU to feed it data across the PCIE bus.
F@H has difficulty with large primes and their multiples number of CPUs.
7 is always large, 5 is sometimes large, and 3 is never large. Try to choose a number that is a multiple of 2 and/or 3.
2, 3, 4, 6, 8, 9, 12, 16, 18, 24, 27, etc. are good numbers of CPUs to choose.
5. 10. 15, 20 etc may work most of the time. Other numbers will bite you

Type the number you want, and click save.

Adjusting the threads down to 3 and putting in on full has worked well, so thanks for the suggestion. CPU temps are currently plateaued in the mid 70s C. GPU temp has remained at 64C. I'm still going to try and mess with airflow, because I feel like I shouldn't be seeing the temps I am with the setup I have. I do regularly blow the dust out of the case, and did so today just to make sure that wasn't the issue. Appreciate the suggestions.

Folding will use a lot more processing power than most games which use hardly any cores.

That's actually a relatively small cooler for that CPU. Are you running a super compact build? I'd put at least a DH-14 dual tower on that cpu. Or a small 120mm all in one liquid cooler or a 240mm if there's room in your case.
I would pull that cooler off and check your tim spread though just in case you didn't use enough.

ipkh wrote:That's actually a relatively small cooler for that CPU. Are you running a super compact build? I'd put at least a DH-14 dual tower on that cpu. Or a small 120mm all in one liquid cooler or a 240mm if there's room in your case.
I would pull that cooler off and check your tim spread though just in case you didn't use enough.

It's a mid size tower, unfortunately, so I don't have a whole lot more room for a larger cooler. Anything wider (dual fan, etc.) would likely overhang my RAM slots, and anything taller is likely to be pressed right up against the side of the case.

FYI, your processor is rated at 82W TDP and max temp of 74C. JimboPalmer's suggestion will work, although you will see temperature spikes depending on your monitoring software poll rate. I'm working off of a laptop, so my cooling capacity isn't as robust as yours and my temps for my i7-8750H were running at around 90-95C with all threads enabled (default setting of FAH). My processor is 45W TDP, Max temp 100C. Other than limiting the cores available, there is another way to approach it, and it is perhaps a little more effective. At full load, my processor frequency is 4 Ghz. When I limit the number of FAH cores, the frequency will still spike up and give me temperature spikes of 95-98C. But upon further investigation, I located some processor settings in the Windows power management settings. So if you go to settings>power options>edit plan settings, you will see a link for "change advanced power settings". Click on that and it will open up a dialog box tree menu. In that menu you will find "Processor Power Management", open that up and in that sub-menu there is a setting for "Maximum processor state". In my case, reducing this to 95% limited my CPU frequency to 2.0 Ghz and brought my temperature down to 73C, which is certainly better than 95C. Even better, it eliminated the frequency swings altogether. The frequency was effectively locked at 2.0 Ghz and the temperature did not fluctuate at all. Oh, my GPU is a GTX 1070 mobile. TDP 145W. It sucks up a lot of power and puts out a lot of heat. Usually sits at around 73C fully loaded. I'm getting some pretty tasty WU's that yield 100K+ points. CPU WU's are only worth about 2-5k by so I'm throttling the CPU in favor of the GPU.

Hope this helps!

Alienware 17 R5
i7-8750H (8th Gen Coffee Lake)
GTX 1070 M
32GB 2666Mhz RAM
Windows 10

The Noctua compatibility list says that cooler has only moderate overclocking headroom with that CPU. I just read an article in Maximum PC that some CPUs can [temporarily] pull as much as 4x TDP when automatic overclocking kicks in...

I have the NH-D14 cooler in both my rigs. The radiators are clear of the RAM slots. The second fan is optional, and can be adjusted to fit over the top of medium-tall RAM sticks. You could measure your rig and see if it would fit. Alternatively, an all-in-one liquid cooler, with the radiator replacing your top fans, may work better.

A couple things you can try:

Disable or limit the automatic overclocking in BIOS if you can. The base clock of 4 GHz is still quite fast.

Try cleaning and re-applying the thermal paste again. When you remove the cooler this time, look to see how well it spread last time, and adjust the size of the drop accordingly. Remember that too much is as bad as too little (especially if it overflows the CPU case, into the slot).

If the CPU cooling fan is somewhat aligned with the case exhaust fan, ensure the CPU fan is blowing directly toward that exhaust.

If the backplane slot covers on either side of your GPU are solid, replace them with slotted covers. This will help direct hot air around the GPU directly out of the case. Though the GPU card backplane may be slotted, few cards actually direct the majority of cooling air through their backplane. It may be counter-intuitive, but also try shutting down one or both of the top exhaust fans. If you keep positive pressure (more intake than exhaust airflow), allowing the hot GPU air to exhaust naturally through the backplane may be a better solution.

I have exactly the same CPU (i7-4790K) and until recently was using the stock cooler. (GPU is RTX2080.)

With the stock cooler and running F@H or R@H, it would hit 100C and throttle down by about ~15%. In R@H for extended time would cause total shutdown. I did some research and found this particular line of Intel CPUs is notorious for having bad heat transfer internally. (According to Tom's Hardware at least.) So in order to keep it in operating range at full load it needs abnormally large cooling solutions.

I picked out the PH-TC14PE based on Tom's Hardware test results and it's working out pretty well. The clock is running at 4.18 GHz with no throttling. It stays at around 65-70c (GPU-Z) with F@H set to 7 cpu threads and 1 gpu thread. But the heat sink / fan just barely fits in the case. One of the fans is not in the ideal place due to DIMM clearance issues. But I'm happy with the numbers. To get any better results at this point it would be easier to just buy better hardware.

It kicks off enough heat that keeping the room its self cool starts to become a problem. Hot air is tending to stagnate around the case which can drive the temperature up. Something to think about is the airflow around/in/out of the room to give the cooler better air to work with.

WussyDan wrote: I have a Noctua U12S 120mm single fan tower on the CPU

Can you add a second fan to that cooler and run it as push/pull ? I took a quick look at Amazon and it looked like it might be possible ?

foonix wrote:I have exactly the same CPU (i7-4790K) and until recently was using the stock cooler. (GPU is RTX2080.)

It kicks off enough heat that keeping the room its self cool starts to become a problem. Hot air is tending to stagnate around the case which can drive the temperature up. Something to think about is the airflow around/in/out of the room to give the cooler better air to work with.

Yeah, so what I've done in the past is open the door in the room I'm working in and placed a small fan in the doorway blowing cool air into the room with it pointed parallel to the floor. Doing this pushes cool air along the floor and the warmer air leaves through the top of the doorway...a natural thermocline. But if you want something that is even more effective, you can significantly lower your case temperature by opening up or removing the side panel and using the fan to blow air directly into the case. That works really well. Only thing better is a water-cooling set up.

Does anyone have any experience with the U12A? After some rough measurements, I'm reasonably certain I won't be able to fit a bigger dual tower like the D14/15 or PH-TC14PE in the case, but from everything I've read the U12A seems like a pretty significant (if not inexpensive) upgrade.

Salient other questions raised, the cooler fan is roughly level with the exhaust fan; currently it's set to pull air from the front intake fans over the cooling tower and back toward the exhaust fan. The backplate entirely open/slotted.

WussyDan wrote:Salient other questions raised, the cooler fan is roughly level with the exhaust fan; currently it's set to pull air from the front intake fans over the cooling tower and back toward the exhaust fan. The backplate entirely open/slotted.

Next easy step is to control the backplane. Except for the slots immediately on either side of the GPU backplane, block those slots with tape. You want to focus the airflow around the GPU to exit the case.