High CPU temps

Topic

New Reply

There is something screwy going on with my processor recently.

I use CoreTemp in the tray to monitor core temps. It used to read in the high 30’s to mid 40’sC with occasional bursts into the 60’sC. Then a few weeks back I started seeing temps soar into the 70-80’sC frequently and even hit 100C at a few points.

I checked in with HWiNFO64 and was also seeing the same numbers as CoreTemp.

I thought maybe a CPU fan had died (I have water cooling on the CPU) but when I looked at the fans, they were both running (sandwiching the radiator). The output hose felt fairly warm but the return hose was cool, so the radiator was working. So I decided to clean the dust out of the box and pull the water cooler on the CPU to look at the thermal paste as I haven’t looked at this in the ~7 years since I originally built this system.

I blew out a lot of dust and the thermal paste coverage looked bad. I cleaned the old paste and then put new paste on and reseated the water cooler.

This helped a bit and brought the temps down but they are still higher than I think they should be (in the high 40’s to mid-50’sC generally. Even when I wake the system up from an overnight sleep, the temps still show high 30’s to mid-40’s.

This was CoreTemp temperatures a few days back when I returned to the system and woke it from deep sleep:

This was HWiNFO temp data on the same day (later).

Right now, CoreTemp is showing this and the CPU is reading about 97% idle

So with all that preamble done, the question is what is going on? How do I figure out what the problem is?  Or if there is a problem at all?

– Looking at the 3rd CPU that got thermal throttled in the HWiNFO attachment, is it possible that the CPU is going bad? Or at least CPU ID 2?

– If I am returning to a system that has been sleeping all night, shouldn’t I be seeing CPU temps close to ambient (around 20°C)?

Reply | Quote

Replies

Have you looked at Task Manager to see what is using the CPU?

You could have a Crypto Miner Virus or something similar?

May the Forces of good computing be with you!

RG

PowerShell & VBA Rule!
Computer Specs

Reply | Quote

No, it is just a mix of regular and system tasks running.

Reply | Quote

Hmmm… You’ve already done what I would have done, so no more ideas yet. Have you ruled out a task constantly running with a high CPU load?

BTW, if you meant to attach anything, I don’t see it.

That’s certainly a weird one! Will keep watching this topic because I am interested to see what is actually going on here… 😉

Windows 10 Pro 22H2

Reply | Quote

I would expect high 40s even with water cooling because the fans should run slowly until required – I assume they are controlled by the motherboard.

cheers, Paul

p.s. having fans sandwiching the radiator is one fan too many. Each fan will only pass as much air as its blades will allow. 2 fans does not equal twice the air.

Reply | Quote

“p.s. having fans sandwiching the radiator is one fan too many. Each fan will only pass as much air as its blades will allow. 2 fans does not equal twice the air”

This is the recommended configuration from the manufacture.  One fan sucks air out of the box and  forces it into the radiator where it picks the radiated heat.  The 2nd fan exhausts the warmed air from the radiator into the room.

Is your statement based on something you read somewhere?

Reply | Quote

It’s basic physics. Fan A can pass X amount of air when running at speed Y. If you attempt to force more air through the fan it is no longer doing any work and is potentially limiting the airflow.
To move more air you either need to run the fan faster or use a larger fan. This is why we use larger fans at slower speeds to achieve the same cooling with less noise.

cheers, Paul

Reply | Quote

“I would expect high 40s even with water cooling because the fans should run slowly until required – I assume they are controlled by the motherboard.”

Why?  If the machine is sleeping, there should not be very little CPU activity, other than regular polling checking to see if it needs to awaken the machine, therefore the temp of the CPU’s should be near ambient, which is around 20°C.

As for the fans, I have control through a mobo app and have tuned them to increase on a specific curve as CPU temp increases.

I could force the CPU rest temps a few degrees lower by speeding up the fans but at the expense of higher noise, which I am not looking for.  However, the rest temps would still be far above ambient.

Reply | Quote

ibe98765 wrote:

Why? If the machine is sleeping, there should not be very little CPU activity, other than regular polling checking to see if it needs to awaken the machine, therefore the temp of the CPU’s should be near ambient, which is around 20°C.

Unless the PC is completely shutdown, the CPU, even in sleep mode, is still drawing some power so it’ll never be at ambient!

For most CPU’s, the “idle” temp range shouldn’t be more than about 20° C above ambient. If it is, then something’s up with either the fans, the AIO cooler (if you have one) or the thermal paste (even the best thermal paste will start to lose heat transfer efficiency if it’s old enough.)

Reply | Quote

There is something screwy going on with my processor recently.

I use CoreTemp in the tray to monitor core temps. It used to read in the high 30’s to mid 40’sC with occasional bursts into the 60’sC. Then a few weeks back I started seeing temps soar into the 70-80’sC frequently and even hit 100C at a few points.

I checked in with HWiNFO64 and was also seeing the same numbers as CoreTemp.

I thought maybe a CPU fan had died (I have water cooling on the CPU) but when I looked at the fans, they were both running (sandwiching the radiator). The output hose felt fairly warm but the return hose was cool, so the radiator was working. So I decided to clean the dust out of the box and pull the water cooler on the CPU to look at the thermal paste as I haven’t looked at this in the ~7 years since I originally built this system.

I blew out a lot of dust and the thermal paste coverage looked bad. I cleaned the old paste and then put new paste on and reseated the water cooler.

This helped a bit and brought the temps down but they are still higher than I think they should be (in the high 40’s to mid-50’sC generally. Even when I wake the system up from an overnight sleep, the temps still show high 30’s to mid-40’s.

This was CoreTemp temperatures a few days back when I returned to the system and woke it from deep sleep:

This was HWiNFO temp data on the same day (later).

Right now, CoreTemp is showing this and the CPU is reading about 97% idle

So with all that preamble done, the question is what is going on? How do I figure out what the problem is?  Or if there is a problem at all?

– Looking at the 3rd CPU that got thermal throttled in the HWiNFO attachment, is it possible that the CPU is going bad? Or at least CPU ID 2?

– If I am returning to a system that has been sleeping all night, shouldn’t I be seeing CPU temps close to ambient (around 20°C)?

Reply | Quote

Once I reposted my OP to try and get the attachments to appear, the attachments now started showing up in the original post!  Whew.

Can a moderator clear up this mess?

Reply | Quote

Have you looked at Task Manager to see what is using the CPU?

You reply:

No, it is just a mix of regular and system tasks running.

It’s not clear if you sorted the Task Manager columns by “CPU” to see what is causing high usage.  You probably did, but it’s hard to image CPU temps so high without high CPU usage and a sort would quickly show the problem.

Sysinternals “Process Explorer” would give you a deeper dive on CPU usage.

Desktop Asus TUF X299 Mark 1, CPU: Intel Core i7-7820X Skylake-X 8-Core 3.6 GHz, RAM: 32GB, GPU: Nvidia GTX 1050 Ti 4GB. Display: Four 27" 1080p screens 2 over 2 quad.

Reply | Quote

TechTango wrote:

Have you looked at Task Manager to see what is using the CPU?

You reply:

No, it is just a mix of regular and system tasks running.

It’s not clear if you sorted the Task Manager columns by “CPU” to see what is causing high usage.  You probably did, but it’s hard to image CPU temps so high without high CPU usage and a sort would quickly show the problem.

Sysinternals “Process Explorer” would give you a deeper dive on CPU usage.

Here is a current example from a few minutes ago.  You can see that the system is at idle yet CPU temps are reading in the high 50’s to low 60’sC.  This should be causing my fans to ramp up their speed, but that isn’t happening.

Yet the temp numbers are the same in two different monitoring apps (CoreTemp & HWiNFO).

Something isn’t making sense here.

Reply | Quote

Just by coincidence, I was checking my CPU temps and Task Manager here while running at idle, and noticed that Windows Shell Infrastructure Host (sihost.exe) was running constantly at 20% CPU with my CPU temps in the mid-40’s range. My PC normally idles in the mid-30’s with negligible CPU usage.

So I rebooted and things are back to normal. This is the first time that I have noticed this bug.

It’s covered here: https://www.makeuseof.com/windows-shell-infrastructure-host-process-high-resources-consumption/

Windows 10 Pro 22H2

Reply | Quote

TechTango wrote:

Have you looked at Task Manager to see what is using the CPU?

You reply:

No, it is just a mix of regular and system tasks running.

It’s not clear if you sorted the Task Manager columns by “CPU” to see what is causing high usage.  You probably did, but it’s hard to image CPU temps so high without high CPU usage and a sort would quickly show the problem.

Sysinternals “Process Explorer” would give you a deeper dive on CPU usage.

I did and I do.

1 user thanked author for this post.

TechTango

Reply | Quote

@ibe98765 –

Ok, so your radiator is doing it’s job and you’ve replaced the thermal paste between the chip itself and the heat sink. The CPU is running barely above idle with about 0-3 percent usage at idle.

The fan on top of the heat sink needs to be spinning at the right speed for a given temperature, OR the pump within your liquid cooling solution needs to be doing the same, running at a given speed for a given CPU temperature. This setting is usually contained within the BIOS/UEFI settings for CPU thermal management, and might even have a fancy name as well. Could it be that for some unknown reason, the setting got changed by a wayward software update that might have recently been installed for your overall thermal solution that includes the radiator that handles the liquid cooling?

Go into your UEFI/BIOS settings to see what the settings are for the thermal management of your CPU, and see if one (or more) has (have) been changed from what might be a perfectly acceptable factory default.

One other thing…see if the tubing/piping for the liquid is damaged in any way that would be slowing down the liquid’s flow. As alluded to above, liquid flowing slower than it should will increase your CPU’s temperature because it won’t be able to remove as much heat as liquid that’s moving faster.

Reply | Quote

[ibe98765]

There is no fan “on top” of the water cool unit.  There is a pump that circulates the water but I have no external control over the pump speed/volume.

As mentioned the output hose is fairly warm.  The return hose is cool, as it should be.  There is no evident of damage to the water tubes.  Heat transfer appears to be effective.

You can see in the screenshot below the current state of the system.  Note that the CPU speed is showing running at 4400 MHZ, even though the system is at 97% idle, which doesn’t make any sense.  If accurate, this indicates the CPU is running fast for no reason.  Doing so would generate excess heat.

If I watch the CPU speed on this screen, it jumps around quite a bit (I assume the system is reacting to minute needs) but it does spend a good deal of time at 4000 Mhz or above.

• This reply was modified 1 year, 10 months ago by ibe98765. Reason: Hit submit too soon

Reply | Quote

What’s the CPU cooling set for: Control Panel\All Control Panel Items\Power Options\Edit Plan Settings ?

Reply | Quote

0%, Active, 100%

Reply | Quote

There is a forum where they use different coolers and fans, try asking there.  https://www.overclockers.com/forums/forums/cooling.6/

Do you happen to have screenshots of the lower idle temperatures from the past, just so nobody questions the idea that it once was lower?

My first thought is that CPU at idle can mean a whole lot of different “states” or actions.  Depending on the power profile you have chosen in Windows, idle could go to a low frequency and voltage partial sleep mode, or it could stay at normal frequency and voltage.  If anything could have changed your power profile to high performance instead of balanced or power saving, I wonder if that could be related.  Power settings are per user, so if you created a new user account that could change it.

If it isn’t something like that, I wonder if the fans are failing not by running slow, but by running hot.  Does your monitoring show that the fans are running at the expected RPM?  If they are standard parts they may be very cheap to replace compared to something like a CPU or even a radiator.

Reply | Quote

I haven’t been to OC forum in some time but I believe I still have an ID there.  I’ll check them out.

No earlier screenshots that I have saved of CPU temps.  Obviously, I would expect others to have the lower temps I describe for this CPU and an idling system, so I shouldn’t have to prove anything.

Yes, there is a lot going on with CPU tuning.  I have played with the BCLK Frequency and CPU ratio just for kicks but doing so usually winds up with a locked system and I have to restore the BIOS.  Tools that I have found related to the CPU seem to be high level and basically show speeds or temp.  I haven’t found any kind of detailed CPU state tracing app yet.

The fans are fine.  They run quiet and I have set fan profiles so that they increase speed as CPU temp increases.

I could throttle down the CPU and blast the fans (noisily) to full RPM and likely obtain some degree of temp reduction but this wouldn’t be pleasant and shouldn’t be necessary.

Again, the point here is that the CPU has recently begun running around 20°C hotter (based on what the CPU temp monitoring programs are showing) than previous (around 60°C as compared to ~40°C before) with no changes made to the system.  Looking at the CPU-Z program, I can see that this processor (rated at 4.0 MHz) regularly jumps into the 4100-4400 range, which should be in overclock territory but I am not doing OC to the best of my knowledge.  So it shouldn’t be going up that high.  I need to figure out why it is doing that for no good reason as this system is just mostly cruising.  I don’t game.  I basically just use it for web surfing and some light occasional image editing).

Reply | Quote

Per the specs from Intel, your processor is designed to operate between 4 and 4.4 GHz, which is equivalent to 4000 and 4400MHz. So, it’s operating within its designed specs. The specs can be found on the following Intel page:

https://ark.intel.com/content/www/us/en/ark/products/80807/intel-core-i74790k-processor-8m-cache-up-to-4-40-ghz.html

Many terms on that page have a ? mark inside a circle next to the term. Clicking on that will help define what’s meant by the term.

Despite the temperature difference that you can feel on the pipes on each side of the heat sink, I still think that just maybe the pump is running more slowly than it should be. Is there any way you can look up the specs for your cooling system to see if they say how fast the pump is supposed to run (in RPM)? One thing that would make it run more slowly would be if it’s not getting the full voltage it’s supposed to be getting.

Reply | Quote

I have no visibility nor control over the workings of the water pump.  It screws onto the CPU and that’s it.

Reply | Quote

I spent a bunch of time on Tuesday reviewing the BIOS settings and playing with fan tuning.

I found a setting in the BIOS that seemed to allow the CPU to go above 4.0MHz, so disabled that and it worked to keep the CPU at below 4.0MHz, which reduces the PCU heat a bit.  There is no work being done on this system to go into “turbo” mode.

Then I changed the CPU fans profile to speed them up about 1/3.  This makes the machine noiser (and a bit annoying) but did significantly drop the CPU temps.  Temps dropped about 15-20°C at system idle.

I still don’t understand how upping the airflow volume while the water flow volume stays the same causes such a big temp decline.

Now I want to figure out why idle temps are not closer to ambient temps (20°C) at idle and when I return to the system after it has been in deep sleep for some number of hours.

Seems like no one else here pays attention to CPU temps!

Reply | Quote

As I pointed out in my above post #2555828, it’s impossible for a working CPU to be at ambient temps because, even at full idle/sleep (i.e. doing nothing), it’s still using some power and power = heat!

As for why increasing the air flow lowered your CPU temps, the explanation is quite simple.

The temperature of the air “inside the case” directly impacts how hot the CPU gets (i.e. the hotter the temp inside the case, the hotter the CPU will be.) Adding more cool outside air into the case by speeding up the fans lowers the “inside the case” air temps which lowers the CPU temps.

This effect is normal for all PC’s, regardless of whether you’re using fan cooling or water cooling for the CPU.

A consequence of this is, CPU’s tend to run hotter if located in an area with higher room temps and lower if the room temps are cooler (which is why servers are normally placed in “temperature controlled” environments with room temps ~18° C.)

ibe98765 wrote:

Seems like no one else here pays attention to CPU temps!

Nope, HWiNFO64 is always running while my PC is active and I “regularly” check the sensors!

Just for reference, here’s my system specs:

• Asus Rog Maximus XI Gene motherboard
• Intel i7 9700K (OC to 4900 MHz)
• G.skill Ripjaws V 32 GB RAM (OC to 4000 MHz)
• Nvidia RTX 3060 Ti graphics card
• Corsair H60 120mm AIO for CPU cooling
• 4 Corsair ML-120 PWM 120mm case fans (max RPM 2500)

My normal idle CPU temps are 28 – 33° C with fan speeds in the 650 – 850 RPM range and an ambient room temp of 21 – 22° C.

My normal full load CPU temps are 69 – 72° C with fan speeds in the 1800 – 2000 RPM range and the same ambient room temp.

Note: idle temps are when the CPU isn’t doing anything, full load temps are when I’m processing video files.

Here’s a screen cap of my current sensor readings while viewing AskWoody.

You’ll notice under the cooling section toward the bottom, the sensors show the RPM for the Corsair H60 pump even though, like yours, it simply attaches directly to the CPU. That’s because the “power” for the pump is provided by the motherboard and it can sense its RPM.

Unless it’s using an “external” power source, I would expect your pump’s RPM to also be available via a motherboard sensor.

To see if it is, click the Status button at the bottom of the SIV64X window and take a look at all the fans displayed in the lower right section.

If one of them has a very high RPM, that’ll be your pump (on my system the pump shows up as item 7 AUX Fan 4 with an RPM above 4,000.)

If none of them have such a high RPM, the either your pump isn’t being powered by the motherboard or it’s not running as fast as it should.

BTW, exactly what make/model is your water cooler?

2 users thanked author for this post.

ibe98765, satrow

Reply | Quote

Just another Forum Poster wrote:

As I pointed out in my above post #2555828, it’s impossible for a working CPU to be at ambient temps because, even at full idle/sleep (i.e. doing nothing), it’s still using some power and power = heat!

Yes, but HOW MUCH heat is generated? You appear to be contending that an idling CPU is generating 20°C of heat.  I seriously doubt that is anywhere near accurate, even in my old CPU.

Just another Forum Poster wrote:

The temperature of the air “inside the case” directly impacts how hot the CPU gets (i.e. the hotter the temp inside the case, the hotter the CPU will be.) Adding more cool outside air into the case by speeding up the fans lowers the “inside the case” air temps which lowers the CPU temps. This effect is normal for all PC’s, regardless of whether you’re using fan cooling or water cooling for the CPU.

I only increased the speed of the CPU fans (that’s one fan pushing air into the radiator and 1 fan exhausting that air to the outside).  The inside of the case is not much different in temps as you can see by noting on my post #2555896 that the mobo temp was 26°C and after increasing the CPU fans speeds was 24°C in post #2556191.

Just another Forum Poster wrote:

My normal idle CPU temps are 28 – 33° C with fan speeds in the 650 – 850 RPM range and an ambient room temp of 21 – 22° C.

Your CPU is newer than mine and uses less wattage, so would naturally run cooler.  We need to compare oranges to oranges.

Re: the water pump – I fired up HWiNFO but don’t have the same info that you show.  Possibly because my mobo is older?

I am going to have to crawl under the desk and open the box later because I am unsure where I have the water pump connected to.  I have an unused header named PWR, which might be where I should have had the pump plugged into to get RPM reports.

Reply | Quote

ibe98765 wrote:

Your CPU is newer than mine and uses less wattage, so would naturally run cooler. We need to compare oranges to oranges.

Except for some special laptop CPU’s that are deliberately designed to use less power, a newer CPU won’t necessarily use less power than an older one! How much power a particular CPU will use depends on its TDP (Thermal Design Power) rating, mine is 95W while yours is only 88W, and what frequency it’s running at, mine is running at 4900 MHz while yours is now running at 4000 MHz.

Remember, regardless of how new or old a CPU is, higher frequency = needs more power = hotter CPU.

BTW, your SIV64X screen shot shows your CPU pulling 15W where mine shows my CPU pulling 30W; which means yours “should” be running cooler than mine (exactly what I’d expect since mine is overclocked from the stock 3900 MHz up to 4900 MHz!)

ibe98765 wrote:

Re: the water pump – I fired up HWiNFO but don’t have the same info that you show. Possibly because my mobo is older?

Probably not!

I “customized” the sensor panel display for my HWiNFO so it only shows the items I was interested in monitoring (i.e. I “hid” all the other items) and I also “changed” the text used for each item to make it easier to ID exactly what it was (i.e. my pump originally displayed as “AIO pump” but I changed the text so it displays as “CPU (Corsair H60 Pump)“.

FYI, whichever header you have the pump plugged into on the motherboard would be the one to check in HWiNFO for the RPM, or as I suggested above, simply use the Status button on SIV64X to find it.

1 user thanked author for this post.

ibe98765

Reply | Quote

Just another Forum Poster wrote:

BTW, your SIV64X screen shot shows your CPU pulling 15W where mine shows my CPU pulling 30W; which means yours “should” be running cooler than mine (exactly what I’d expect since mine is overclocked from the stock 3900 MHz up to 4900 MHz!)

Exactly.  And I am not.  Which was the crux of the problem.

Anyway, I dove into the hardware and discovered (for reasons lost in the fog of memory), that I had the CPU1 & 2 fans plugged into a Y connector into CPU fan header #1.  And I had the water pump plugged into CPU fan header #2.  So I reconfigured this to what it should be with the 2 CPU fans plugged into the proper CPU fan headers and the water pump plugged into the PWR header.

After doing this, the water pump started making a lot of noise, when it used to be rock silent quiet plugged into CPU fan header #2.  Don’t know why this started happening!  But possibly the water pump wasn’t working very well because there wasn’t enough power for the pump from the CPU fan header?  Therefore it would be running slower (and quieter) but not doing a good job of cooling, which is why I was seeing the very high CPU temps.

Right now, temps are staying where they used to be at idle, which is in the mid-high 30’sC:

So I ordered a new water cool unit from Amazon that will arrive tomorrow.

It was tough to find a 120mm radiator unit that fits in my old case and mounts on the internal rear.  Only 3 or 4 choices and only 2 that could arrive tomorrow.  Most modern units seem to have two or three 120/140mm fans that mount on the top.  While I have room for two 120mm fans on top, it didn’t look like the the radiator and fans together would clear the memory slots and also, the PWR connector is up on the high right side of the mobo, so would also be blocked.

Reply | Quote

Paul T wrote:

It was tough to find a 120mm radiator unit that fits in my old case and mounts on the internal rear.

I feel your pain!

My case has the same issue were the radiator must be mounted at the rear and there’s absolutely no room for a larger AIO. I dread the day I have to replace my 120mm Corsair H60 with another 120mm AIO which, as you discovered, are getting very hard to find!

In fact. I’m tempted to order a replacement now so I’d have it “on-hand” when that day finally arrives but… see below about how water levels change over time in older AIO’s.

ibe98765 wrote:

While I have room for two 120mm fans on top, it didn’t look like the the radiator and fans together would clear the memory slots and also, the PWR connector is up on the high right side of the mobo, so would also be blocked.

Ran into exactly the same problem when trying to upgrade the 120mm AIO cooling in my nephew’s gaming PC with a 240mm AIO!

It turned out the extra drive slots in the front of his case were “removable” and, since they weren’t being used, we took them out and put a 240mm radiator and its cooling fans in the front of the case which dropped his “gaming temps” back down to a more reasonable 75 – 80° C from the previous 80 – 85° C.

ibe98765 wrote:

After doing this, the water pump started making a lot of noise, when it used to be rock silent quiet plugged into CPU fan header #2. Don’t know why this started happening! But possibly the water pump wasn’t working very well because there wasn’t enough power for the pump from the CPU fan header?

Since it was plugged into a fan header, the system most likely saw it as just another fan so your fan curves were controlling how much power it got and thus how fast it ran.

BTW, depending on how old your AIO is, it’s possible the water level inside has gotten lower and the noise you’re hearing is air bubbles circulating thru the pump. This happens because, even though it’s a “sealed system” some water still manages to very slowly evaporate out of it (typically via the hoses/hose connection points) thus lowering the total amount of water remaining in the system.

This is normal for all water cooling systems, both AIO and external, because water is notorious for being able to find even the smallest spot were it can leak out of a container!

Another thing that can cause a pump to suddenly get really noisy is if an air bubble (all AIO’s have “some” air in them) gets trapped either in the CPU heatsink block or at the IN/OUT hoses to the radiator. To avoid this possibility the top of the radiator (the end opposite the IN/OUT hoses) should be the “highest” point in the system (air will always rise to the top most point in a closed cooling system.)

For more info about this with some really good “visual” demonstrations of the effect, see Stephen Burke’s Stop Doing It Wrong: How to Kill Your CPU Cooler video.

1 user thanked author for this post.

ibe98765

Reply | Quote

This indicates the pump was running too slowly and the noise is from it now running at full speed.

Ideally you want the pump to run faster as the CPU needs more cooling, so having it connected to a fan header would seem to be the correct method, as long as you can set the base speed independently on each fan header.

As a test, put the pump back on the fan connector and increase the minimum fan speed.

cheers, Paul

Reply | Quote

Just another Forum Poster wrote:

Ran into exactly the same problem when trying to upgrade the 120mm AIO cooling in my nephew’s gaming PC with a 240mm AIO! It turned out the extra drive slots in the front of his case were “removable” and, since they weren’t being used, we took them out and put a 240mm radiator and its cooling fans in the front of the case which dropped his “gaming temps” back down to a more reasonable 75 – 80° C from the previous 80 – 85° C.

Whew. Those are pretty high temps!

How about an external radiator cooling setup?

Reply | Quote

ibe98765 wrote:

Whew. Those are pretty high temps!

Actually, since he’s running an i7-10700K overclocked to 5200 MHz and likes to run his games with all the setting maxed out, those 80 – 85° C temps were well within the expected range for his overclocked CPU. He just wanted to try and get them even lower because, the lower the max temp the CPU runs at, the longer everything will last before it might fail.

ibe98765 wrote:

How about an external radiator cooling setup?

Unlike my own case, his doesn’t include any “knock-outs” for external cooling hose connectors so that wasn’t really an option; without either purchasing a new case or major modifications to his existing case.

Since the temps he was seeing weren’t really outside the range some gamer’s rigs reach (even 85 – 90° C is pretty common), a larger AIO was a much simpler solution.

Reply | Quote

Just another Forum Poster wrote:

BTW, depending on how old your AIO is, it’s possible the water level inside has gotten lower and the noise you’re hearing is air bubbles circulating thru the pump. This happens because, even though it’s a “sealed system” some water still manages to very slowly evaporate out of it (typically via the hoses/hose connection points) thus lowering the total amount of water remaining in the system. This is normal for all water cooling systems, both AIO and external, because water is notorious for being able to find even the smallest spot were it can leak out of a container! Another thing that can cause a pump to suddenly get really noisy is if an air bubble (all AIO’s have “some” air in them) gets trapped either in the CPU heatsink block or at the IN/OUT hoses to the radiator. To avoid this possibility the top of the radiator (the end opposite the IN/OUT hoses) should be the “highest” point in the system (air will always rise to the top most point in a closed cooling system.)

Mine is set-up this way.  Seems like there should be a gauge to see the water level in all systems.  And maybe a bleed valve.  I assume the more expensive refillable ones have this but I haven’t seen it mentioned on the sealed units.

Reply | Quote

Just another Forum Poster wrote:

FYI, whichever header you have the pump plugged into on the motherboard would be the one to check in HWiNFO for the RPM, or as I suggested above, simply use the Status button on SIV64X to find it.

It is fantastic how much info is available through the free HWiNFO and SIV!  Someone here should write an article on these utilities to make others aware.

Reply | Quote

Hey Y’all,

I’m finding this discussion very interesting. Having never been an over-clocker, with the exception of my IBM-PC-AT (changing the 6Mhz crystal for an 8Mhz one), but I’ve always been interested in the topic.

I had never given any thought to changing the thermal paste, but now I’ve seen it mentioned here and on a couple of LTT videos. So I though I’d check and see how my geriatric computers were doing.

Purchased Manufacturer CPU
07/15/13 Dell137000 LT i5-5200U
05/15/15 DellXPS8700 DT i7-4470
08/05/17 DellXPS8920 DT i7-7700

We’ll go from oldest to newest:

Dell137000

DellXPS8700

DellXPS8920

Note: Ambient temp. in the room is 74F.

Well if Dell is doing nothing else right, it looks like they were doing a good job with the thermal paste.

May the Forces of good computing be with you!

RG

PowerShell & VBA Rule!
Computer Specs

Reply | Quote

Most everyone uses Centigrade temps when OC’ing.  It’s hard to compare Fahrenheit temps.

Anyway, unless you pull your CPU’s and replace the existing paste with a high-end paste, which might be warranted for machines 5 or more years old and regularly used, you won’t know how good a job a Dell is doing.

Reply | Quote

I’m finding this discussion very interesting. Having never been an over-clocker, with the exception of my IBM-PC-AT (changing the 6Mhz crystal for an 8Mhz one), but I’ve always been interested in the topic.

I’m posting this as a reference for an overclocked CPU running Windows 10.   Overclock percentage is 29% on Asus MB.  One click overclocking using the “fast” setting, not the higher “fastest”.

I have no instabilities on the “fast” setting.  I did have instabilities with the “fastest”

07/2018 Home built desktop Asus MB Intel Core i7-7820X Skylake-X 8-Core 3.6 GHz. Overclocked 25%. (Liquid cooling.  Standard thermal paste)

This CPU is one of the few 7th gen chips approved for Win 11.

Desktop Asus TUF X299 Mark 1, CPU: Intel Core i7-7820X Skylake-X 8-Core 3.6 GHz, RAM: 32GB, GPU: Nvidia GTX 1050 Ti 4GB. Display: Four 27" 1080p screens 2 over 2 quad.

Reply | Quote

ibe98765 wrote:

Most everyone uses Centigrade temps when OC’ing.  It’s hard to compare Fahrenheit temps.

Huh? Max temp is 212F. High recorded 145F. I don’t see a problem making the comparison.

I’m also not going to tear apart a working computer that’s showing no signs of problems.

May the Forces of good computing be with you!

RG

PowerShell & VBA Rule!
Computer Specs

Reply | Quote

RetiredGeek wrote:

Huh? Max temp is 212F. High recorded 145F. I don’t see a problem making the comparison.

I believe he was actually referring to making a comparison with other’s results in Celsius without having to do a F/C conversion.

The majority of user posts on the CPU topic that I have seen elsewhere use Celsius as the international standard of temperature measurement. Fahrenheit is mostly a US only standard, with a few minor exceptions. https://en.wikipedia.org/wiki/Celsius

But use whatever suits you and allows you to manage your machine. That’s why there is a toggle in the CPU monitor program! However it’s easier to make comparisons with other users if everyone shares a common standard of measurement.

Windows 10 Pro 22H2

2 users thanked author for this post.

Zig, ibe98765

Reply | Quote

One thing I discovered with this noisy water pump is that some process wakes up my system overnight but then when the process is completed, the Win10 machine does not go back to sleep.  So I wake in the morning to a loud water pump spinning away.

Acronis backup overnight could be one process that wakes the computer.  But of course there is the all too frequent running ‘Host process for Windows Services’ (a very dumb identifier that does not tell me WHAT process is actually being serviced) and numerous others.

Before I go to bed at night, I click the machine into sleep mode manually, which basically suspends everything in around 15 seconds.

Does anyone have any ideas on how I might diagnose who/what the problem is in this situation?

Is there a trace program I can run that shows what runs minute by minute?

Reply | Quote

ibe98765 wrote:

But of course there is the all too frequent running ‘Host process for Windows Services’ (a very dumb identifier that does not tell me WHAT process is actually being serviced)…

One way to put a positive ID on just what is being serviced is the built-in-to Windows utility of Task Manager.

Right click on the task bar and select “Task Manager” from the resulting list. Now that you’ve selected it, if it’s the first time you’ve ever run it, click on the item labeled “More details” in the lower left corner of the window. Now, click on the tab towards the top of the window that’s labeled “Processes”. You”ll now be looking at a list of processes that are running. Those that are using the “svchost” to be run will show up as “Service Host: Name of service being run by svchost”, where the text after the colon will actually name the service that svchost is actually running. To see what the last CPU resource hog is, click the top of the column labeled “CPU”. There should now be a little “down arrow” next to the listing for “CPU”. This puts the running processes in order from those taking the most of the CPU’s time (or clock cycles) to those taking the least of the CPU’s time (or clock cycles).

Doing the above procedure is but one way to see just what is waking the system up at night, but you have to be awake to actually see it. I’m confident there are a few third party utilities that will take this info and make a graph of just what’s been going on and when, so you don’t have to be awake to see it at that very moment, you can see it at your own leisure. 😉

BTW, have you installed the new cooling system you ordered from Amazon yet?

Reply | Quote

ibe98765 wrote:

some process wakes up my system overnight

Open an Admin Command Prompt (Win R, cmd, Ctrl Shift Enter)
Type: powercfg /waketimers

Let us know what you have.

cheers, Paul

Reply | Quote

Paul T wrote:

ibe98765 wrote:

some process wakes up my system overnight

Open an Admin Command Prompt (Win R, cmd, Ctrl Shift Enter)
Type: powercfg /waketimers

Let us know what you have.

cheers, Paul

C:\>powercfg /waketimers
Timer set by [SERVICE] \Device\HarddiskVolume2\Program Files (x86)\Common Files\Acronis\Schedule2\schedul2.exe (AcrSch2Svc) expires at 3:59:00 AM on 5/16/2023.

Timer set by [SERVICE] \Device\HarddiskVolume2\Windows\System32\svchost.exe (SystemEventsBroker) expires at 2:29:31 AM on 5/16/2023.
Reason: Windows will execute ‘NT TASK\Paragon Job (Back up partitions or disks to virtual containers) – {470b9658-7d68-402d-1fc9-96b7f95dfc7c}’ scheduled task that requested waking the computer.

Hmmm.  So does look like Acronis wakes the computer to run.  However, I forgot about the Paragon task that I have been playing with in the past and will remove this.

Still, the question remains as to WHY the system doesn’t return to sleep after the job is done.

Reply | Quote

ibe98765 wrote:

…Still, the question remains as to WHY the system doesn’t return to sleep after the job is done.

I haven’t used the sleep or hibernate function for a great many years after having severe problems with it back in XP. I decided it wasn’t worth the hassle, so since then I’ve completely disabled both sleep and hibernation on all of my machines since then.

However, to take a stab at giving you a possible answer to your question…perhaps the machine needs to be put back to sleep manually or by a specific instruction set that needs to be invoked after being awakened by a task within Task Scheduler.

Just a guess.

Good to hear that your new cooler has seemingly solved your problem of excessive CPU heat. I believe that @bigal67 (alejr) hit the nail on the head when he said that all liquid cooling systems, even those touted as being sealed, develop very minute leaks over time.

That type of leak is what I believe led to a drop in the water cooling level and subsequently introduced one or more air pockets that made noise as they traveled through your system once you put the pump on a different bus with a higher voltage which made the pump spin a lot faster.

The core problem is resolved, so you can set the resolved flag with your next reply to anyone in this thread if you’d like. BUT, you can also leave the status just as it is until you get a concrete answer to your other question regarding getting the computer to go back to sleep after the task(s) that awoke it has/have finished.  😉

Reply | Quote

Bob99 wrote:

Doing the above procedure is but one way to see just what is waking the system up at night, but you have to be awake to actually see it. I’m confident there are a few third party utilities that will take this info and make a graph of just what’s been going on and when, so you don’t have to be awake to see it at that very moment, you can see it at your own leisure.

BTW, have you installed the new cooling system you ordered from Amazon yet?

Yes, but no one here has suggested such a utility yet.  Wondering if I might be able to see this info in the event logs?  Will have to look.

Yes, I have installed the new CoolerMaster water unit (https://www.amazon.com/dp/B086C1PK2Y).  It is very quiet even though the pump spins at 2400RPM (800 more than the old one) and works much better than the old ThermalTake unit.   Below are some sample temps on the 4 CPU’s upon waking the system yesterday.  As can be seen, they are reasonably close to ambient, which is what I would expect.

1 user thanked author for this post.

Bob99

Reply | Quote

Temps looking REALLY good! My i5-9500 NOT overclocked by any means is currently running 28C while typing this, and doing nothing else. BTW, it’s conventionally cooled by the cooling solution that came in the processor’s box from Intel. Can’t even hear the cooling fan it’s running so slowly! However, I do hear it when I start an anti-crapware scan with Defender or, on rare occasion, Malwarebytes.

Reply | Quote

ibe98765 wrote:

Yes, I have installed the new CoolerMaster water unit (https://www.amazon.com/dp/B086C1PK2Y). It is very quiet even though the pump spins at 2400RPM (800 more than the old one) and works much better than the old ThermalTake unit.

That seems to be a nice AIO at a reasonable price, but I think I’ll go with the NZXT Kraken 120 (https://www.amazon.com/NZXT-Kraken-120-RL-KR120-B1-Effective/dp/B09GKQ4KYP) when mine finally needs to be replaced.

It’s only $16 more and is the newer version of the NZXT M22 which consistently gets rated higher than the CoolerMaster in online AIO comparisons of 120mm AIO’s (mainly due to it’s 6 yr warranty and faster pump speed.)

Reply | Quote

ibe98765 wrote:

WHY the system doesn’t return to sleep

Your backup utility may have an option to put the machine to sleep after the backup has run – mine does.

cheers, Paul

Reply | Quote

Paul T wrote:

ibe98765 wrote:

WHY the system doesn’t return to sleep

Your backup utility may have an option to put the machine to sleep after the backup has run – mine does.

cheers, Paul

Just checked, I don’t see an option like that.  Will send it to suggestions.

What I do have are the following two options:

1. Wake up the sleeping/hibernating computer
2. Prevent the computer from going to sleep/hibernate

They were both checked on.  I unchecked the latter under the theory that perhaps this wasn’t applying only while the backup was running.

Reply | Quote

Is there an option to run a program post backup? If so you can send the computer to sleep via the command.

PSShutdown from Sysinternals is easy.
psshutdown.exe -d -t 0 -accepteula

cheers, Paul

1 user thanked author for this post.

ibe98765

Reply | Quote

OK, that seems to have worked.  Thanks!

I’ve had the sysinternals and NirSoft programs installed for a long time but because there are so many, I have rarely looked through them.  I need to sit down and spend some time doing this.

The freeware WSCC program is a very useful UI for accessing and keeping these programs up-to-date.  Someone here should do a write-up on WSCC.

Reply | Quote

Why don’t you do the write up and we will add it to the newsletter?

cheers, Paul

Reply | Quote

One crazy thing I am experiencing since adding that sleep command to the Acronis backup tasks is that every day when I return to my computer, between 10:00am and 11:00am, my system goes into sleep mode while I am using it!

The Acronis tasks have been run hours before and therefore should not be triggering the command.

I have searched the scheduled tasks to see if the sleep command somehow got added in there but see nothing.  The only clue I have is that whenever this happens, the following event record shows just before the system went into sleep.

Log Name: System
Source: e1dexpress
Date: 6/20/2023 10:41:56 AM
Event ID: 27
Task Category: None
Level: Warning
Keywords: Classic
User: N/A
Computer: pc-03222015
Description:
Intel(R) Ethernet Connection (2) I218-V #2
Network link is disconnected.

I cannot figure out what network link it is referring to or how to control it.

I’ve thought of removing the sleep command I added to the Acronis backups to test if it is just coincidence between adding the sleep command and running into this problem but that is a very blunt approach.

Any other ideas anyone?

Reply | Quote

You need to remove the sleep command from the backup to find out if that is the problem. You only need to do this for a couple of days.

cheers, Paul

Reply | Quote

Problem solved.

Windows 10 Power control has always been screwed up.  For example, even though I have a password entry required when returning from the lock screen, this only works when I physically click ‘Winkey-L’  If I let the system go into [supposed] lockdown by inaction timeout, hitting any input will reopen the system w/o a password being requested.  It doesn’t matter much because I live alone, but still…

In this case, I enabled the following settings in the Acronis app:

Advanced settings
[X] Back up only when the computer is locked or screensaver is running
[X] Wake up the sleeping/hibernating computer
[X] Prevent the computer from going to sleep/hibernate

I had #’s 2 & 3 originally checked on.  Adding #1 seems to have fixed the issue.

Without this checked on, the backup jobs were running when I returned to the system in the morning and then when completed, were issuing the sleep command.  #2 checked on should have addressed this but apparently did not.

Reply | Quote

ibe98765 wrote:

What I do have are the following two options: Wake up the sleeping/hibernating computer Prevent the computer from going to sleep/hibernate They were both checked on.  I unchecked the latter under the theory that perhaps this wasn’t applying only while the backup was running.

Unchecking the latter option did not make any difference.

Reply | Quote

Paul T wrote:

Is there an option to run a program post backup? If so you can send the computer to sleep via the command. PSShutdown from Sysinternals is easy. psshutdown.exe -d -t 0 -accepteula

How do you run a program from something like Sysinternals?

I have sysinternals programs as part of the WSCC suite but how do I invoke one indirectly if I didn’t first start the program to receive commands?

Reply | Quote

You can run (almost) any program by entering it in a Command Prompt, as shown in my example – don’t forget to add the path to the front of the command.
Search C: for “pssutdown” to find the full path.

cheers, Paul

Reply | Quote

Question resolved.

Reply | Quote