Intel Core i5 4690 frequency drops to 800 MHz ingame

[Distroyer]

Hey,
as stated in the topic my CPU goes down to 800 MHz even while gaming.

It first happened about 2 weeks ago and I don't know what caused it. After that I updated the BIOS, reinstalled most drivers and nvidia driver, because I didn't know the low fps were caused by the CPU. But that didn't work either and I reinstalled Windows 10 64Bit.

Now every now and then the CPU speed goes back up to 3.65 - 3.8 GHz and stays there a few minutes if I'm lucky.

My Hardware is the following:
Mainboard: MSI Z97 Gaming 7
CPU: Intel Core i5 4690 4x 3.50 GHz (not overclockable)
Graphics Card: MSI GeForce GTX 960 Gaming 4G
RAM: 8GB Corsair Vengeance DDR3-1600

I read about a Slow Mode switch on the Mainboard and tried to figure out if it's on or off by flipping it and reboot, but it was off at the beginning and the speed got worse then flipped. So that's not the problem.

Next I checked the temperature of the CPU on full load while stressing and it went slowly up to 66°C, which shouldn't be a problem either. On idle it's ~34-40°C.

I encountered the problem mostly in Heroes of the Storm (Blizzard), where 800 Mhz seems to cause 1-2 FPS and increases your latency, and in Rocket League, where the Frames drop not that much but still noticable, Sound gets laggy and played twice, and gameplay gets laggy too, even if the ingame information shows normal 40-60 ms.

So in conclusion:
- BIOS updated
- Windows reinstalled
- Drivers updated
- CPU not overheating
- Slow Mode switch is off

I got the Hardware June 2015 and it worked until about 2 weeks ago.
I hope somebody knows what the problem is or can help me, because 1-2 FPS are not fun. 🙁

Edit:
I use CPU-Z and Prime95 for stress-testing and core-temp for the temperature.

Edit2:
Forgot about Windows Energy Options. Minimum and Maximum were set to 100%, then the CPU speed jumped between 3.68 MHz and 0.77 MHz.
Now I set Minimum to 20% and now it's also between these frequencies.

 

[GhislainG]

Install Intel XTU, select all monitors, set it for one hour and then stress test the CPU. Does the processor frequency drop during the stress test? Is thermal throttling always 0%?

 

[Distroyer]

Ok, the test is now running for 5 minutes and the CPU utilization went up to 95-100% as expected, the Core Frequency is at 0.80 GHz all the time, Package Temperature is at 34-36°C, but Thermal Throttling is at 100% all the time.

Does that mean, the cpu is slowing down because it or the mainboard thinks it's too hot although the Package Temperature is at 34-36°C?

Edit:
25 minutes later, no change.

 

[GhislainG]

What are you using to stress test the CPU? The XTU stress test? Thermal throttling should be 0%; what are the core temperatures?

 

[Distroyer]

Now I'm using XTU, the Core Frequency jumped up to 3.70 GHz for a few seconds and so did the Temperature increase to max 60°C. The Thermal Throttling jumped down to 0% when that happened, so I assume that allowed the Processor to increase the Speed. The core temperatures are about the same as the Package Temperature, ~37°C at 0.80 GHz.

Is it a Sensor Error on the Mainboard? I found several Posts with the same problems.

 

[GhislainG]

Does it run at a constant 3.70 GHz when using the XTU stress test (not really a stress test)? My i5-4590 CPU temperature peaks up to 69°C on that stress test and it obviously doesn't throttle. Throttling should not occur until the CPU or core temperatures reach 100°C (unless there's a setting in the BIOS to throttle the CPU at much lower temperatures). Temperatures at idle are meaningless.

 

[Distroyer]

As long as it doesn't throttle it tries to max out at 3.70 GHz.
I checked the BIOS for any options related to CPU throttling, but couldn't find anything suspicious.

 

[GhislainG]

What are the temperatures just before it throttles? What's nice with XTU is that you can hover the graphics or and see the values at any given time.

 

[Distroyer]

I have to rerun the test, but it was at the 36°C and as the cpu got faster the temperature was going up to about 60°C. There were no temperature spikes or anything. The core temperature wasn't a very low or high temperature at all the time beside the normal increase because of the speed increase.

Also I have to be lucky to get a unthrottled moment in the test.^^

Edit: Not getting a full speed time, but before the throttling goes up to 100% again, the temperature reached one time 47°C and wasn't increasing fast and in the past it reached higher temperatures (max ~60°C) before it gets lowered by the throttle.

 

[GhislainG]

It definitely looks like the motherboard is throttling the CPU; it could be a motherboard sensor issue. You should contact MSI support and provide your observations while monitoring with XTU. The CPU can't get hot if it doesn't run at full speed and it has no reason to throttle if it stays below 100°C.

Since you can't overclcok your CPU, I presume that you never used OC Genie? You may want to read this: https://www.reddit.com/r/buildapc/comments/2w0pg7/troubleshootingnew_processor_i74790k_running_at/ Is there an Eco mode in the BIOS?

 

[Distroyer]

I think the Eco switch is the Slow Mode switch, that I already checked. I guess there really isn't much I can do about it. Thank you very much for your support! I will try contacting the MSI support.

Edit:
Here are 4 screenshots from XTU.
The computer just booted after several hours and the Throttle was luckily at 0% but is at 100% now.

 

[Distroyer]

I replaced the MSI Z97 Gaming 7 with a new Mainboard, Gigabyte GA-H97-HD3, and ran the XTU Stress Test.
The Thermal Throttling is constantly at 0% and Package Temperature reaches about 70°C and the Core Frequency is at 3.69GHz. So it really seems to be an issue with a sensor on the Mainboard.

 

[GhislainG]

Probably, but you could ask MSI what was wrong (they may not provide that info).