I know, I know, but i figure I throw out my issue here as well.
For a while i have noticed my GPU or CPU is throttling due to overheat in Windows when running relatively modern games.
During my trial and error I found out that
I had neglecting cleaning and the CPU, GPU and intake fans were very dusty. - Fixed.
It’s a pre-built tower and the intake fan was placed awkwardly. I moved it a bit and it now gives a more powerful airflow directly on the GPU and CPU (before it was aiming to first cool the HDD, but I don’t need it to do that). - Fixed
I tested to put a big ol’ tower fan outside of the computer just blowing air into the intake from 40 cm away (not with an open chassi! just in front of the normal intake) and suddenly i could up the resolution and run smoothly even. So DEFINITELY a heating issue.
Finally today my wife was out of town so I could game with speakes instead of headphones… and realized what the core issue is: The fan control isn’t working. All fans except the GPU who handles it’s own fan are always idling, aka on the lowest specced RPM.
The fans are not BROKEN, they spin and during bios check at startup all fans sound like a hurricane, as always. So even the speed controller itself works. It’s just that in Windows it no longer registers the need to increase speed at higher temperatures.
I have ordered a new gaming rig that should be here in a month or so (damn tech shortages) but in the meantime I have ordered a 14 cm chassi fan to put in along the other intake fan, and this one is primitive, it will always run at max speed…
In the meantime tho I did something I haven’t done since the early 2000s: I broke loose the next plate in the back next to the GPU card slot, meaning now the GPU fan can throw out about 35% of it’s heated air directly thru that slot 2 cm from it. It seems to have helped, monitoring my temp while playing today the GPU never went over 82C while before it was up over 98C (!!). Same settings.
So… My question is: Anyone else have experienced this and in that case what caused it?
I don’t, this is a pre-built system that has worked since 2019. And as I said when the boot check happens at start all the fans go full speed in 3 seconds like they always have. I have not checked in Linux because well… I have no games installed there (here). Sometime during the summer it just… happened.
Oh I know. That said, as I said above… this system has been able to run 1080p / 60 Hz on high or medium settings every new-ish game I have thrown at it and all old ones (say Fallout 4, Sims 4, Skyrim) on Ultra for two years. The cooling was obviously not optimal but it stayed workable for two years, I will take apart the new one when it arrives mainly to move the HDD in this one in there as a media storage unit but I will check fan placement while I’m at it.
Already done. I will try a BIOS defaults reset just to be sure, but it seems to stay cool enough to be workable with the extra slot opened up in the back and the intake fan moved. Adding an extra intake fan will definitely help, and then I can survive for a month or so
OKay, so… It is probably something shot on the motherboard. The Speed Control works, as evident by the fans speeding up when the BIOS self check starts at boot, and the Temp Sensors also work as evident by well… the computer being able to display a visibly varying temperature of the CPU, the GPU and the HDD when asked to, both in Linux and Windows.
However, the connection to actually speed the fans up because of the increased temperature seems to be broken both in Windows and Linux, so… weird (There is no speed controls in the BIOS, and I did a BIOS reset).
Oh well, lesson learned: Double check the airflow (as I said just opening that slot in the back behind the GPU decreased the average gaming temperature with 15C (!!!) which is a heck of a lot. At least the new gaming rig will be better cooled (it has 5 fans built in instead of two, and the card is a 3060, so it’s not over the top but able to handle more than I can throw at it (I only have a 1080p / 75Hz monitor and do not plan to buy a new one, so the GPU won’t be pushed to it’s limits ever). I will see if I can open a slot in the back for it to start with anyway…
Is there an application running in the background that controls fans and their speeds?
I know I have, as I have an Asus Rog Strix motherboard, and the application that’s installed for it can control the fan profile automatically or manually, and when they ramp up and down again. I can set it to kick the fans into different speeds dependent on the temperature of the GPU/CPU, so it could be it’s set the ramp to never happen.
No program, no separate setting in bios. It just detects the fans in bios (correctly I might ad) but there is no option to control them other than the normal power scheme settings (prioritize performance over cooling or the other way around so windows know if it should speed up the fans instead of throttling the CPU)
Not sure what’s going on here. Oh well, added a second intake fan, had to do some improv rigging since I went for an 14 cm fan in a space made for a 12 cm fan (deliberately; I had a plan and it worked).
On the other hand all this shows that it’s definitely worth opening up your pre-built tower and check airflow.
With very minor adjustments I got my old workhorse NVIDIA 1050 TI to idle in Windows at 32 degrees(!) and in Linux at 40 degrees (Linux, on all machines i have ran NVIDIA drivers on which so far is two laptops and one tower, always ran hotter than in Windows doing 2D stuff).
What’s more important: I can now run Forza Horizon 4 on 1080p High settings, 75fps and that old card fluctuates between 77 (!!!) and 85(!) degrees, most of the time staying on 82 degrees (which is damn good, especially for an old card like that)
Before these adjustments, even with the fans speeding up a lot, the temperature was > 95 degrees,