[SOLVED] CPU throttling when gaming, temperature is fine

tyranius

Distinguished
Feb 15, 2012
38
0
18,530
So, for the past month I've been having this problem where the CPU will downclock during gaming. The intervals vary between every 10 to 45 minutes. On average it happens every 25 minutes or so. The CPU will be running at 3.8GHz as usual then downclock to 2.6GHz for a couple of seconds, then come back up. This happens even when the cores are at 40ºC.

After some googling, the ideas I have are, VRM temperature and some kind of PSU problem. My PSU is 10 years old [Corsair VX550], my PC itself is about 5 years old. Unfortunately, my mobo does not have VRM sensors, so I haven't been able to specify whether VRM temps are the issue. Also, could it be the PSU? It's really old, but from what I've read, PSUs that can't meet power demands would cause severe system instability. My PC doesn't crash or hang or anything. It simply downclocks the CPU during gaming. Other than that, the thing is tip-top.

So, is there anything I can do to isolate and identify what the problem is? I always keep the mobo pretty clean, which makes me kind of skeptical about the VRM idea. The thing is that it's the only culprit of CPU throttling I could find that was unrelated to CPU temperature.

Help?

The specs:

CPU: A10-7860K
GPU: 750Ti
PSU: Corsair VX550
Mobo: A58M-A/BR
 
Solution
You kind of answered your own question here. In lighter workloads it does not need to run at 3.8 ghz so it down-clocks automatically. Your more intensive games require it to run at 100% so it does. All CPUs do this for efficiency and to conserve energy. If its running at 3.8 all the time and doesn't need to the hardware will degrade much faster.

tyranius

Distinguished
Feb 15, 2012
38
0
18,530
I stress-tested the CPU for about 10 minutes, the CPU reached 55ºC, but did not throttle to 2.6 as usual, it stayed at 3.8 throughout the test. Makes me think this is not a mobo VRM temp issue. The games I've had this happen generally have low core usage. League of Legends, PUBG Lite, Bloodstained... it did happen in CS:GO too though.

I'm starting to think this is a problem of the CPU thinking it should be idle, then downclocking, reaching 100 usage and going "oh <Mod Edit>, my cores are actually in use" then going back up and repeat. Next step is to disable Cool n' Quiet in the BIOS and see if that changes anything. If nobody posts any other ideas I think I'll just format everything as a last resort.
 
Last edited by a moderator:

M3rKn

Respectable
Nov 13, 2019
315
70
1,890
What stress test did you run? 10 minutes is not enough, realistically 15 minutes is a bare minimum, 30 minutes to an hour would be ideal. Since your game is not crashing I don't see much of an issue, CPUs throttle for various reasons. It could be some background process requiring CPU usage. You mentioned your PSU is 10 years old. A weak or inferior PSU can cause issues, but at 10 years old it has well outlived its usefulness. I would replace it for sure.

edit: if its been more than 2 years since you re-formatted then it would actually be a good idea. Its just normal maintenance.
 
Last edited:

tyranius

Distinguished
Feb 15, 2012
38
0
18,530
Thanks for the response.

I used HeavyLoad stress test. Ran it for an hour and same thing. CPU gets a little hot, like 55ºC but no throttling happens. I also fired up Hitman[2016] because that game literally keeps my cores at 100% at pretty much all times and the GPU at 90-100%. Played for an hour with no issues either.

So I went and changed the windows power plan to Performance, which makes the CPU run at 100% at all times. I've been playing Bloodstained for 2 hours and monitoring with HWINFO, no throttling as of yet. For some reason, my CPU just throttles while playing CPU light games, but with games that use it to its full capacity, it never throttles. So it seems my initial suspicion was correct, while playing light games my CPU thinks it can just go idle.

While running in Performance power plan is a decent workaround, I'd rather not have to keep changing the power plan every time I boot up a game. With the most likely culprit out in the open, maybe someone has some other suggestions to fix this problem. Why does my CPU think it can go downclock while obviously running an application?
 

M3rKn

Respectable
Nov 13, 2019
315
70
1,890
You kind of answered your own question here. In lighter workloads it does not need to run at 3.8 ghz so it down-clocks automatically. Your more intensive games require it to run at 100% so it does. All CPUs do this for efficiency and to conserve energy. If its running at 3.8 all the time and doesn't need to the hardware will degrade much faster.
 
Solution