icue doesnt react to stress tesr

[YoungVegabond]

hi all as i've stated in the title i have a problem with my corsair h150 elite cappelix where during a stress test the tempa remain pretty much the same. its important to notice that i also had this issue with my previous aio (corsair h150 pro xt). i'll really appreciate your help.

Edited February 21, 2023 by YoungVegabond

[LeDoyen]

what's happening exactly? what test are you running, for how long? what CPU do you have? what temperatures? 🙂

[YoungVegabond]

34 minutes ago, LeDoyen said:

what's happening exactly? what test are you running, for how long? what CPU do you have? what temperatures? 🙂

hi i ran xtu and cinebench cpu stress tests  for 5 minutes. i have i9 10850k and my temp on idle are 25-35 amd during stress test the max i ever reaches is abou 70c..

[LeDoyen]

it works fine if the CPU stays that cold.

You may need to run the stress test for longer for the water to stabilize. Depending on your fan curve, it may take longer to reach steady state. what water temperatures do you get at idle and under load when the water temp stops climbing?

[YoungVegabond]

44 minutes ago, LeDoyen said:

it works fine if the CPU stays that cold.

You may need to run the stress test for longer for the water to stabilize. Depending on your fan curve, it may take longer to reach steady state. what water temperatures do you get at idle and under load when the water temp stops climbing?

under load 32c and wile idle 27-28c. and unfortunatly i ran the tests a lot of times but nothing helped.  i've had this problem even with the previous aio (h150 pro xt).

[YoungVegabond]

even if it works fine its a serious problem if i cant see my real time temps there...

[Corsair Notepad]

When you are running the tests to replicate the issue on iCUE are you also running other hardware monitoring tools? If not now exactly do you know the temps you are experiencing then if the iCUE software isn't reporting them?

If you are the likely cause of the monitoring not working on iCUE is a mixture of iCUE being put on a lower CPU priority while the highly CPU intensive benchmarks run meaning iCUE will be delayed on polling and reporting the information. And also the other hardware monitoring software preventing iCUE from being able to read and report the sensors.

[YoungVegabond]

9 minutes ago, Corsair Notepad said:

When you are running the tests to replicate the issue on iCUE are you also running other hardware monitoring tools? If not now exactly do you know the temps you are experiencing then if the iCUE software isn't reporting them?

If you are the likely cause of the monitoring not working on iCUE is a mixture of being put on a lower CPU priority while the highly CPU intensive benchmarks run meaning iCUE will be delayed on polling and reporting the information, but also the other hardware monitoring software preventing iCUE from being able to read and report the sensors.

thanks for the reply. i used hawinfo couple of times but i read its bad for icue so i stopped and only used cinebench or xtu ever since. its also important to notice that the issue started after i replaced my pc case at the store but i checked the connections and everything seems fine and even after replacing the entire aio its still present..

[YoungVegabond]

i really dont know what's the cause of this.

[LeDoyen]

Could it be that your new PC case just has way better airflow than the previous one and you don't see as much temperature increase as you used to?

a non-unlocked 10850k (if the power limits are still in place) wouldn't put out much heat tbh.

Edited February 22, 2023 by LeDoyen

[YoungVegabond]

1 hour ago, LeDoyen said:

Could it be that your new PC case just has way better airflow than the previous one and you don't see as much temperature increase as you used to?

a non-unlocked 10850k (if the power limits are still in place) wouldn't put out much heat tbh.

but even during stress tests it remains constant and doesnt change..even thoogh xtu shows the temp to be near 70c, its really weird..

Edited February 22, 2023 by YoungVegabond

[LeDoyen]

aaah then it's not the water temp that doesn't change. it's just iCUE monitoring that doesn't reports the CPU temperature correctly when you run a stress test, keeping it flat?

As long as the AIO is set to regulate fan speed based off water temp, it will not depend on iCUE to give it readings, so you should be safe there. Now for the CPU temp not changing, that would be an iCUE dashboard problem and not an AIO one. I have seen a few occurrences of that a few weeks back in the forum but didn't see if there was a solution for it.

[YoungVegabond]

41 minutes ago, LeDoyen said:

aaah then it's not the water temp that doesn't change. it's just iCUE monitoring that doesn't reports the CPU temperature correctly when you run a stress test, keeping it flat?

As long as the AIO is set to regulate fan speed based off water temp, it will not depend on iCUE to give it readings, so you should be safe there. Now for the CPU temp not changing, that would be an iCUE dashboard problem and not an AIO one. I have seen a few occurrences of that a few weeks back in the forum but didn't see if there was a solution for it.

unfortunatly i'm searching more than a year for a solution but still didnt find anything that helps.  and from what i'm seeing the cpu temp does change its just that the fans speed and aio temp remain the same (at least according to the icue) and wont change during stress test.  sorry for the confusion english is not my native language.

Edited February 22, 2023 by YoungVegabond

[c-attack]

On 2/21/2023 at 6:38 AM, YoungVegabond said:

i ran xtu and cinebench cpu stress tests  for 5 minutes. i have i9 10850k

 

On 2/21/2023 at 8:15 AM, YoungVegabond said:

under load 32c and wile idle 27-28c. and unfortunatly i ran the tests a lot of times but nothing helped.  i've had this problem even with the previous aio (h150 pro xt).

It doesn’t look like there is a problem. This how cpu cooling works. If you start a stress test and the cpu jumps from its 30C idle to 70C, then 70C is the lowest possible temp you can ever have at that specific room/case temperature. Your cpu is cooled conductively regardless of cooler type and if I give you an air box cooler, a 360mm AIO, and a massive external custom cooling array, your cpu temp after 1 second of the stress test for all is 70C. That is the temp you get when you put X volts through that exact cpu. 
 

Where the cooling design and type comes with waste heat removal. They all transport that conducted heat elsewhere. But the more watts you put in, the more watts in needs to press out or that liquid temp will increase and it adds to the cpu temp. The cooling type and efficiency determines how much worse the cpu temp will get as the load continues over time. Quit after 20 seconds and they all look the same. Run for 30 min and there are clear differences. The air cooler will start a slow and steady cpu temp rise after 5-10 seconds and may increase another 15-25C. It can’t hold a lot of heat although really large air boxes can be pretty effective. A 360mm radiator will go up about +10C with a 300W load and fans around 1300 rpm. Most CPUs output far less, so +5C for XTU and 150W seems about right. If you had an enormous wall sized radiator and multiple pumps, you could dissipate every watt added in a single pass and the cpu liquid temp would never increase. You would stay at 70C indefinitely, while the 360 warms to 75C, and the little air box may run away to 85-90C. This plus the noise of the fans is what differentiates various cooler types and sizes. 
 

So if your coolant is only going up 5C during the test, then it is only adding 5C cpu temp max. That also means you can only reduce the current cpu temp by 5C at some unbelievably high fan speed. In reality +3C is about as low as you can go on most AIO units and none of this matters for short XTU or R20 single pass tests. If you were to run a longer test (like R23) you would want more fan speed. At present the fans are not speeding up because you are down near 30C water temp. The curve is hidden from you but there is an X:Y temp:RPM curve. 
 

Click on the yellow + bar in the cooling tab to create a new custom curve. A graph will appear below. Mouse over to under the graph and those 4 shape tools. Three of them correspond to the Quiet, Balanced, and Extreme preset but they will show their points when you click them. Pick any one you want, name it whatever you want, and you also can move points. If you want the fans to run at 1300 rpm at 35C, you can do this. However, do make sure you change the sensor to H150 Elite Cap Temp. Using cpu temp is inefficient, creates a lot of dynamic fan change noise, and won’t help your temps one bit. Remember the fans can’t keep the cpu temp from spiking to 70C. They do keep that number from rising when continuous load is applied. 
 

Also note the custom curve open the fixed speed option. I suggest you use this for stress testing. You can pick any speed you want, but it should be consistent from test to test for comparison and really makes more sense when testing stability or heat management. Also note none of this will remedy and current CPUID or Windows 10/11 reading issues for cpu load, temp, or Core identification. However, if you’ve set your fans to 1300 or whatever ahead of time, then you don’t need cue and can watch in a detailed monitor window like xtu or hwinfo. 

Edited February 23, 2023 by c-attack

[YoungVegabond]

16 hours ago, c-attack said:

 

It doesn’t look like there is a problem. This how cpu cooling works. If you start a stress test and the cpu jumps from its 30C idle to 70C, then 70C is the lowest possible temp you can ever have at that specific room/case temperature. Your cpu is cooled conductively regardless of cooler type and if I give you an air box cooler, a 360mm AIO, and a massive external custom cooling array, your cpu temp after 1 second of the stress test for all is 70C. That is the temp you get when you put X volts through that exact cpu. 
 

Where the cooling design and type comes with waste heat removal. They all transport that conducted heat elsewhere. But the more watts you put in, the more watts in needs to press out or that liquid temp will increase and it adds to the cpu temp. The cooling type and efficiency determines how much worse the cpu temp will get as the load continues over time. Quit after 20 seconds and they all look the same. Run for 30 min and there are clear differences. The air cooler will start a slow and steady cpu temp rise after 5-10 seconds and may increase another 15-25C. It can’t hold a lot of heat although really large air boxes can be pretty effective. A 360mm radiator will go up about +10C with a 300W load and fans around 1300 rpm. Most CPUs output far less, so +5C for XTU and 150W seems about right. If you had an enormous wall sized radiator and multiple pumps, you could dissipate every watt added in a single pass and the cpu liquid temp would never increase. You would stay at 70C indefinitely, while the 360 warms to 75C, and the little air box may run away to 85-90C. This plus the noise of the fans is what differentiates various cooler types and sizes. 
 

So if your coolant is only going up 5C during the test, then it is only adding 5C cpu temp max. That also means you can only reduce the current cpu temp by 5C at some unbelievably high fan speed. In reality +3C is about as low as you can go on most AIO units and none of this matters for short XTU or R20 single pass tests. If you were to run a longer test (like R23) you would want more fan speed. At present the fans are not speeding up because you are down near 30C water temp. The curve is hidden from you but there is an X:Y temp:RPM curve. 
 

Click on the yellow + bar in the cooling tab to create a new custom curve. A graph will appear below. Mouse over to under the graph and those 4 shape tools. Three of them correspond to the Quiet, Balanced, and Extreme preset but they will show their points when you click them. Pick any one you want, name it whatever you want, and you also can move points. If you want the fans to run at 1300 rpm at 35C, you can do this. However, do make sure you change the sensor to H150 Elite Cap Temp. Using cpu temp is inefficient, creates a lot of dynamic fan change noise, and won’t help your temps one bit. Remember the fans can’t keep the cpu temp from spiking to 70C. They do keep that number from rising when continuous load is applied. 
 

Also note the custom curve open the fixed speed option. I suggest you use this for stress testing. You can pick any speed you want, but it should be consistent from test to test for comparison and really makes more sense when testing stability or heat management. Also note none of this will remedy and current CPUID or Windows 10/11 reading issues for cpu load, temp, or Core identification. However, if you’ve set your fans to 1300 or whatever ahead of time, then you don’t need cue and can watch in a detailed monitor window like xtu or hwinfo. 

thanks for the reply and the detailed explanation! i already have a custom curve, but still have the same issue, the fan speeds and even aio temp remain pretty much the same and wont go up or down despite xtu showing that the cpu temp is more or less 70c.. its been more than a year and i still cant find a solution online.

[c-attack]

As long as you are clear it does not matter whether your CPU temp is 30, 50, or 70C -- the fans react to the coolant temperature.  They can only remove heat from the liquid and so they don't need to speed up or down until the liquid temp changes.  If your coolant temp does not increase, the fans don't increase, and it is not affecting CPU temperature until that liquid temp starts to move up.  

Edited February 23, 2023 by c-attack

[YoungVegabond]

On 2/23/2023 at 2:57 AM, c-attack said:

 

It doesn’t look like there is a problem. This how cpu cooling works. If you start a stress test and the cpu jumps from its 30C idle to 70C, then 70C is the lowest possible temp you can ever have at that specific room/case temperature. Your cpu is cooled conductively regardless of cooler type and if I give you an air box cooler, a 360mm AIO, and a massive external custom cooling array, your cpu temp after 1 second of the stress test for all is 70C. That is the temp you get when you put X volts through that exact cpu. 
 

Where the cooling design and type comes with waste heat removal. They all transport that conducted heat elsewhere. But the more watts you put in, the more watts in needs to press out or that liquid temp will increase and it adds to the cpu temp. The cooling type and efficiency determines how much worse the cpu temp will get as the load continues over time. Quit after 20 seconds and they all look the same. Run for 30 min and there are clear differences. The air cooler will start a slow and steady cpu temp rise after 5-10 seconds and may increase another 15-25C. It can’t hold a lot of heat although really large air boxes can be pretty effective. A 360mm radiator will go up about +10C with a 300W load and fans around 1300 rpm. Most CPUs output far less, so +5C for XTU and 150W seems about right. If you had an enormous wall sized radiator and multiple pumps, you could dissipate every watt added in a single pass and the cpu liquid temp would never increase. You would stay at 70C indefinitely, while the 360 warms to 75C, and the little air box may run away to 85-90C. This plus the noise of the fans is what differentiates various cooler types and sizes. 
 

So if your coolant is only going up 5C during the test, then it is only adding 5C cpu temp max. That also means you can only reduce the current cpu temp by 5C at some unbelievably high fan speed. In reality +3C is about as low as you can go on most AIO units and none of this matters for short XTU or R20 single pass tests. If you were to run a longer test (like R23) you would want more fan speed. At present the fans are not speeding up because you are down near 30C water temp. The curve is hidden from you but there is an X:Y temp:RPM curve. 
 

Click on the yellow + bar in the cooling tab to create a new custom curve. A graph will appear below. Mouse over to under the graph and those 4 shape tools. Three of them correspond to the Quiet, Balanced, and Extreme preset but they will show their points when you click them. Pick any one you want, name it whatever you want, and you also can move points. If you want the fans to run at 1300 rpm at 35C, you can do this. However, do make sure you change the sensor to H150 Elite Cap Temp. Using cpu temp is inefficient, creates a lot of dynamic fan change noise, and won’t help your temps one bit. Remember the fans can’t keep the cpu temp from spiking to 70C. They do keep that number from rising when continuous load is applied. 
 

Also note the custom curve open the fixed speed option. I suggest you use this for stress testing. You can pick any speed you want, but it should be consistent from test to test for comparison and really makes more sense when testing stability or heat management. Also note none of this will remedy and current CPUID or Windows 10/11 reading issues for cpu load, temp, or Core identification. However, if you’ve set your fans to 1300 or whatever ahead of time, then you don’t need cue and can watch in a detailed monitor window like xtu or hwinfo. 

thanks for the reply and the detailed explanation! i already have a custom curve, but still have the same issue, the fan speeds and even aio temp remain pretty much the same and wont go up or down despite xtu showing that the cpu temp is more or less 70c.. its been more than a year and i still cant find a solution online.

thanks a lot man! 🙂

[LeDoyen]

Thing is there's no problem to find a solution for. That's just how watercooling works 🙂