[SOLVED] My CPU is at max clock speed when idle, but goes below base clock speed when under a heavy load.

[thelegendofpyro]

My issue is that when my PC is under a heavy load, the clock speeds actually go below my base clock speed, and when I'm at idle it stays around it's maximum but fluctuates quite a bit. My initial thoughts are that this is because of thermal throttling, but my temperatures are the same as they have always been and my voltage use is the same as it's always been. I'm running an i7-8750h and when I'm gaming I typically stay between 90-95c, and idle is around 40-50c. This CPU is made to run at these temperatures so it is completely normal. However, I use throttlestop to undervolt my PC, and I have a feeling that every time I edit something in Throttlestop it doesn't even change anything because I've played around with it so much. I use the typical settings of -125*c on both cores and cache, and speedshift at 0, that's it. The only thing my task manager shows is System Idle Processes which is irrelevant, and SVChost.exe which takes up practically nothing most of the time. My C0% is relatively high most of the time hovering around 5-10 on idle, and reaching around 20 during heavy loads.

I never had this problem in the past gaming at these levels, and my Throttlestop settings remain the same as of now - my theory is that when I was tweaking with different settings such as the turbo boost power limits in TPL in the past, they somehow stuck after resetting my Throttlestop multiple times by deleting its ini file, and it must have changed something in my bios. Thoughts?

 

[Darkbreeze]

90-95°C is beyond the recommended safe temperature of 80°C (85°C MAX, if we're being very generous). If you think you are thermally compliant with temperatures above that, then nobody here is going to be of any assistance to you, because they know better.

First thing you should do is revise your ideology by reading, and then RE-reading, the Intel temperature guide, and then look at the problem with a freshly revised set of eyes. Then, get rid of throttle stop and all the other BS you have installed, because they aren't helping you, they are making things worse.

You should not need to use ANYTHING other than the default configuration of OS, drivers and current BIOS to have a thermally compliant, stable configuration so long as you are not trying to run other software that is attempting to thwart the intended behavior and boost profiles of the package you are running. I suspect most of your problems are like self caused and the fact you didn't have problems in the past was because no thermal damage had yet occurred. Once it has, that's it. There is no way to reverse thermal damage to a CPU or motherboard. Or most anything for that matter.

I'd probably recommend a clean install, blow the unit out with compressed air if you haven't already. Consider either a disassembly and fresh paste for the CPU heatsink (Or take it to a shop that can do it for you).

I would not be that surprised to see a clean install of Windows and NOT reinstalling throttle stop, to help, but it might already be too late if you've been running that CPU in the 90-100°C range for a period of time.

You will find a lot of misinformation around the web regarding the fact that "laptop CPUs tend to run hotter, and that is fine, they are designed to handle that" which is pure BS. Mobile CPUs from Intel are not any better at handling high temperatures than desktop CPUs from Intel. If it was that, Intel would just design the desktop CPUs to be capable of handling those temperatures without damage as well, so they could ride a higher frequency and tout the package having more performance. But they don't. Because it doesn't work that way. By design, these CPUs were intended, at most, to run in the 80-90 range with temperatures being below 80°C being HIGHLY PREFERRED if you want it to have a nice long lifespan with no cumulative thermal damage that is out of the ordinary.

I'd also recommend looking to see if there is a newer BIOS version, and if there is, install it. If there is not, then I'd probably just reset the BIOS to the default values and then do a clean install of Windows, as a good place to start trying to figure the problem out. And cross your fingers that the problem is not simply due to thermal degradation.

You need to also keep in mind that CPU temps are NOT the only thing that causes thermal throttling, and thermal throttling is EXACTLY what you are dealing with if clock frequency DROPS under a load. VRMs and motherboard temps are a big factor in that as well, and might cause throttling even if CPU temps seem to be, just barely, within tolerance. AND, I'm not saying yours ARE within tolerance either, just that if they were it could STILL throttle if the VRMs are overheating.

 

[thelegendofpyro]

90-95°C is beyond the recommended safe temperature of 80°C (85°C MAX, if we're being very generous). If you think you are thermally compliant with temperatures above that, then nobody here is going to be of any assistance to you, because they know better.

First thing you should do is revise your ideology by reading, and then RE-reading, the Intel temperature guide, and then look at the problem with a freshly revised set of eyes. Then, get rid of throttle stop and all the other BS you have installed, because they aren't helping you, they are making things worse.

You should not need to use ANYTHING other than the default configuration of OS, drivers and current BIOS to have a thermally compliant, stable configuration so long as you are not trying to run other software that is attempting to thwart the intended behavior and boost profiles of the package you are running. I suspect most of your problems are like self caused and the fact you didn't have problems in the past was because no thermal damage had yet occurred. Once it has, that's it. There is no way to reverse thermal damage to a CPU or motherboard. Or most anything for that matter.

I'd probably recommend a clean install, blow the unit out with compressed air if you haven't already. Consider either a disassembly and fresh paste for the CPU heatsink (Or take it to a shop that can do it for you).

I would not be that surprised to see a clean install of Windows and NOT reinstalling throttle stop, to help, but it might already be too late if you've been running that CPU in the 90-100°C range for a period of time.

You will find a lot of misinformation around the web regarding the fact that "laptop CPUs tend to run hotter, and that is fine, they are designed to handle that" which is pure BS. Mobile CPUs from Intel are not any better at handling high temperatures than desktop CPUs from Intel. If it was that, Intel would just design the desktop CPUs to be capable of handling those temperatures without damage as well, so they could ride a higher frequency and tout the package having more performance. But they don't. Because it doesn't work that way. By design, these CPUs were intended, at most, to run in the 80-90 range with temperatures being below 80°C being HIGHLY PREFERRED if you want it to have a nice long lifespan with no cumulative thermal damage that is out of the ordinary.

I'd also recommend looking to see if there is a newer BIOS version, and if there is, install it. If there is not, then I'd probably just reset the BIOS to the default values and then do a clean install of Windows, as a good place to start trying to figure the problem out. And cross your fingers that the problem is not simply due to thermal degradation.

You need to also keep in mind that CPU temps are NOT the only thing that causes thermal throttling, and thermal throttling is EXACTLY what you are dealing with if clock frequency DROPS under a load. VRMs and motherboard temps are a big factor in that as well, and might cause throttling even if CPU temps seem to be, just barely, within tolerance. AND, I'm not saying yours ARE within tolerance either, just that if they were it could STILL throttle if the VRMs are overheating.

I understand that the things next to the CPU can cause higher temps, but clearly you don't know much about this specific model. It is literally notorious for having high temperatures, and there has never been a situation where I've gotten under 90C under a heavy load, which is the reason why I got Throttlestop in the first place. This is just how it is. I don't even think it's possible unless I heavily underperform my CPU which practically renders the laptop useless - which it may be. What I'm looking for is how I can reduce these temperatures without doing that, and why my laptop is now all of a sudden showing signs of thermal throttling despite my temperatures being the same, if not even better than they used to be. It can't be damage done to my parts because that wouldn't make any sense as I've ran my laptop around 100C for years before finding out how to lower those temperatures.

I will see if there's a BIOS update though and try and reset those settings. Thank you.

 

[Darkbreeze]

If you have never been able to run your CPU under a heavy load without it remaining below 90°C, then you have had a faulty laptop/CPU/motherboard/lack of sufficient cooling from the start. It's not normal behavior for ANY Intel CPU to run in the >90°C thermal envelope.

If you think I don't know what I'm talking about, perhaps you'll listen to somebody who happens to be an engineer and has written the Intel temperature guide to put these considerations to sleep. While mobile models aren't specifically covered in the guide last I checked, there has never in the past (Or to date) been an Intel generation where the mobile versions were designed to run in a thermal envelope that exceeds that of same gen desktop parts. The fact that they DO do that is not by design. That is due to poor implementation on the part of the laptop manufacturer who is responsible for frequency, cooling and AVX configurations in the BIOS as well as chassis and cooling designs.

Being a partner for Intel when it comes to buying and selling components doesn't mean that it's always getting done properly by every manufacturer.

And, MORE importantly, if it's done this from the start then I'd have, from the start, recommended taking it back for repair or replacement. There are plenty of precedents showing laptop manufacturers don't always follow the "rules" as outlined by Intel, or AMD, as evidenced by the droves of failed laptops with A8 and A10 processors that mostly all died early thermal deaths due to the inability to keep parts within an acceptable envelope. But, whatever you think. I'm here volunteering to help, not to change your mind or convince you of anything. If you feel it's normal, then more power to you.

I would however suggest you fully read this, which I failed to post before and apologize for that.

How To - Intel CPU Temperature Guide

Update: March 12th, 2024 Preface The topic of processor temperatures can be very confusing. Conflicting opinions based on misconceptions concerning terminology, specifications and testing leaves users uncertain of how to properly check cooling performance. This Guide provides an understanding...

forums.tomshardware.com

And just to offer a little bit of "you might believe that but I'm telling you unequivocally that you, and everybody else on the internet saying those temps are normal are wrong", I'll offer this to you.

The Tjunction temperature for the i7-8750H, as seen in the Intel data sheets, is 100°C. That is not the "safe" operating temperature, that is the temperature at which the system is SUPPOSED to aggressively throttle (Less aggressive throttling CAN happen at temps lower than this, or might not, depending on the laptop manufacturers implementation and hardware table configuration) or shut down if throttle operations are not successful.

https://ark.intel.com/content/www/u...-8750h-processor-9m-cache-up-to-4-10-ghz.html

The 9700k, which is obviously a desktop part, has the SAME Tjunction specification of 100°C, and IS only recommended to be safely run in the 85°C or lower envelope (80°C or lower highly recommended. Once you hit 85°C you are already getting into a probability of thermal degradation if your CPU is consistently "living" there).

https://ark.intel.com/content/www/u...9700k-processor-12m-cache-up-to-4-90-ghz.html


So there is NO "this CPU just runs that hot and it's normal". That's simply not the truth as designed by Intel, and anybody who says otherwise is talking out their earhole.