Jump to content
Corsair Community

How do I setup H100i V2?


Bulent98

Recommended Posts

Hello, I already set up the H100i V2 but I don't think its doing great job or I am just expecting to much stuff from it.

 

I've got i7 6700k @ 4.6GHz and 1.31v

 

When I am stress testing with AIDA64 I get around 78-82 degrees in cores in real temp and when I look at AIDA64 sensors "CPU" I get 60 degrees. Know I don't know exactly which figure is correct but it will be amazing if you guys could help me out.

 

Getting back to the point, got the H100i V2 and the USB is connected as well as the pump is connected to CPU header and I have disabled Q Fan control on my Z170-A PRO MOBO.

 

If you guys are saying the temperatures are fine then that's all good but I got another problem when the temperature goes up sometimes it goes up for a second which is normal as I have read about that but that causes the fans to throttle up too , any solution for that too?

Link to comment
Share on other sites

The AIDA "CPU temp" may not match your core temps. As far as I know, the AIDA monitors a specific temp sensor on the CPU. Other programs may use an average of the core temps that would more closely resemble the core temps themselves. My 5820 and 5930 HW-E also do this at full bore the core temps will be in the 66-72C range while the CPU Temp sensor maxes out in the 55-60 range. A partial explanation may be the temp sensor is on the out shell of the CPU, closer to the cold plate. As such, it stays cooler under max load as most of the heat passes easily through the cold plate. The cores are being heated electrically by the socket. Since they are closer to the heat source, they are warmer. You will usually see the opposite at idle, when the voltage is stepped down the cores are likely to be cooler than the CPU temp sensor which may receive some heat back from the coolant. The exact nature of the values are surely CPU model specific and I can't add any more to that without a 6700K, but in this case it is the core temps that matter.

 

Which AIDA stress test did you run? The interface looks simple enough, but the boxes checked have a substantially effect on the test results. An AIDA FPU test is fairly brutal. A blend test of CPU/FPU/Cache/Memory is generally milder and a CPU only test more so.

Link to comment
Share on other sites

  • Corsair Employee
Hello, I already set up the H100i V2 but I don't think its doing great job or I am just expecting to much stuff from it.

 

I've got i7 6700k @ 4.6GHz and 1.31v

 

When I am stress testing with AIDA64 I get around 78-82 degrees in cores in real temp and when I look at AIDA64 sensors "CPU" I get 60 degrees. Know I don't know exactly which figure is correct but it will be amazing if you guys could help me out.

 

Getting back to the point, got the H100i V2 and the USB is connected as well as the pump is connected to CPU header and I have disabled Q Fan control on my Z170-A PRO MOBO.

 

If you guys are saying the temperatures are fine then that's all good but I got another problem when the temperature goes up sometimes it goes up for a second which is normal as I have read about that but that causes the fans to throttle up too , any solution for that too?

 

Just checking. But do you have the CPU header set to 100%? Just wanting to make sure you're all good on that end.

 

The cores tend to spike a bit while the package stays at a cooler temp. Similar behavior can be seen with the i7-7700K.

Link to comment
Share on other sites

The AIDA "CPU temp" may not match your core temps. As far as I know, the AIDA monitors a specific temp sensor on the CPU. Other programs may use an average of the core temps that would more closely resemble the core temps themselves. My 5820 and 5930 HW-E also do this at full bore the core temps will be in the 66-72C range while the CPU Temp sensor maxes out in the 55-60 range. A partial explanation may be the temp sensor is on the out shell of the CPU, closer to the cold plate. As such, it stays cooler under max load as most of the heat passes easily through the cold plate. The cores are being heated electrically by the socket. Since they are closer to the heat source, they are warmer. You will usually see the opposite at idle, when the voltage is stepped down the cores are likely to be cooler than the CPU temp sensor which may receive some heat back from the coolant. The exact nature of the values are surely CPU model specific and I can't add any more to that without a 6700K, but in this case it is the core temps that matter.

 

Which AIDA stress test did you run? The interface looks simple enough, but the boxes checked have a substantially effect on the test results. An AIDA FPU test is fairly brutal. A blend test of CPU/FPU/Cache/Memory is generally milder and a CPU only test more so.

 

Firstly, thank you for the reply. I ran the normal CPU/FPU/Cache/Memory so basically I left it on default didn't change anything. I also have the overclock voltage set to manual in the BIOS. I didn't do any adaptive offset so even if it's under load I still get 1.31v to my CPU which may cause some high temps when IDLE? But I mean I think its normal but I am getting paranoid I just wanted to make sure that I am getting the temperatures that I am supposed to get.

Link to comment
Share on other sites

Just checking. But do you have the CPU header set to 100%? Just wanting to make sure you're all good on that end.

 

The cores tend to spike a bit while the package stays at a cooler temp. Similar behavior can be seen with the i7-7700K.

 

I have set everything in the bios to PWM 100% all of them then I went and turned off Q fan control to off just to make sure and be in the safe side. That's all I done. This is what I am getting while flying on my flight simulator its quite a CPU intensive simulator CPU usage goes from 60-79.http://imgur.com/a/ZZ1Ev

Link to comment
Share on other sites

I think this is more or less as expected for your voltage and CPU. I also use the CPU/FPU/Cache/Mem 'Blend Test' for my standard stability testing. If you are looking to see if the cooler is working properly, take a look at the starting/ending (load) H100iv2 temperature in C-Link. This is the coolant temperature. You likely idle about 4-7C over your room temperature and I would expect a 6-10C rise during CPU testing. This is both duration and CPU/voltage/wattage dependent.

 

If the coolant temp does not change much at all, it suggests a contact problem between the cold plate and CPU. If changes too much, it usually means some type of flow problem with the cooling system.

 

It also a good idea to run another stability test for a comparison point. OCCT is free and likely to bring back similar numbers. If you want to feel a little better, run Intel XTU. It will likely have lower numbers, but it also useful for troubleshooting. The load is smooth and wave-like, so it becomes easy to see if you messed up your TIM application and get lots of temp spikes. Use the settings wrench on the graph to display cores 1-4 individually and not just 'CPU temp'.

 

EDIT: Also check the peak Vcore in AIDA or whatever other monitoring program you are using after the test. If you are on manual voltage, it should not be overshooting by more than 0.015v or so, but Skylake is famous eating extra voltage beyond the BIOS setting during synthetic testing.

Link to comment
Share on other sites

I think this is more or less as expected for your voltage and CPU. I also use the CPU/FPU/Cache/Mem 'Blend Test' for my standard stability testing. If you are looking to see if the cooler is working properly, take a look at the starting/ending (load) H100iv2 temperature in C-Link. This is the coolant temperature. You likely idle about 4-7C over your room temperature and I would expect a 6-10C rise during CPU testing. This is both duration and CPU/voltage/wattage dependent.

 

If the coolant temp does not change much at all, it suggests a contact problem between the cold plate and CPU. If changes too much, it usually means some type of flow problem with the cooling system.

 

It also a good idea to run another stability test for a comparison point. OCCT is free and likely to bring back similar numbers. If you want to feel a little better, run Intel XTU. It will likely have lower numbers, but it also useful for troubleshooting. The load is smooth and wave-like, so it becomes easy to see if you messed up your TIM application and get lots of temp spikes. Use the settings wrench on the graph to display cores 1-4 individually and not just 'CPU temp'.

 

EDIT: Also check the peak Vcore in AIDA or whatever other monitoring program you are using after the test. If you are on manual voltage, it should not be overshooting by more than 0.015v or so, but Skylake is famous eating extra voltage beyond the BIOS setting during synthetic testing.

 

Thank you for the reply, when you mean the H100i V2 temp is it the one next to the FAN and PUMP RPM? shows 37.7 degrees if so its working flawless because idle it goes around 30 degrees.

Link to comment
Share on other sites

It sounds like the cooler is working effectively. If the core temps during AIDA blend hit 77-82 while the coolant is around 37C, then your difference between coolant temp (lowest possible CPU temp) and actual CPU temp is +40C. That is about where most people run out of room. However, obviously others have been able to run 6700K at 4.8-4.9 GHz with voltages above 1.35v without hitting 90C on the cores. You may not have a golden CPU capable of winning benchmark contests, but when was the last time you changed the TIM? If it has been a while, it might be loosing some effectiveness and this is were is usually evident.
Link to comment
Share on other sites

I built it like 3-4 months ago so it still has the same thermal paste compound I mean I think the temps im getting normal but I was just worried because on the corsair link I get 74 degrees in the package and I got all worried and all that shouldn't be this warm where I am too but these days the weather went from 15 degrees to 30 :/
Link to comment
Share on other sites

An increase in room temperature is additive to the CPU core temperatures, no matter if it is an air tower or water cooler. A +5C increase in room temp will lead to a +5C increase in CPU core temps. You have increased your starting temperature basis. This is why all those benchmark jockeys do their best work in Winter with the windows open.

 

If you can roughly calculate the difference in current temperatures as being equivalent to a change in room temperature, then there isn't much to worry about. For a normal user, 3-4 months is not time to change to TIM yet. Likely only for really heavy users, rendering stations, etc. It is also possible you could steal some degrees back for Summer with a slightly lower voltage. I would think you might be able to get by at 4.6 with a touch less, but as always this is CPU specific.

Link to comment
Share on other sites

An increase in room temperature is additive to the CPU core temperatures, no matter if it is an air tower or water cooler. A +5C increase in room temp will lead to a +5C increase in CPU core temps. You have increased your starting temperature basis. This is why all those benchmark jockeys do their best work in Winter with the windows open.

 

If you can roughly calculate the difference in current temperatures as being equivalent to a change in room temperature, then there isn't much to worry about. For a normal user, 3-4 months is not time to change to TIM yet. Likely only for really heavy users, rendering stations, etc. It is also possible you could steal some degrees back for Summer with a slightly lower voltage. I would think you might be able to get by at 4.6 with a touch less, but as always this is CPU specific.

 

Alright, lastly what do you think of this? http://imgur.com/2nYlnGN

I think its high but end of the day I dont know man it seems like my friends get 60-68 I just get 80 I really dont know anymore.. Hopefully you can help me out. Thank you in advance

Link to comment
Share on other sites

I am open to continuing the look at this. Your friends that run 60-68C... Is it the same stress test, with a H100i v2 (or any other 240mm cooler) and at what voltage?

 

Now that I see the Link data:

 

1) What stress test or program were you running and for how long?

 

2) I can see the end coolant temp is 41C. What was the initial value?

 

3) Your SSD and motherboard temp sensor also read near 40C. This suggests your case ambient temp may also be near the same value. If so, your coolant temp can't be lower than what is and we are looking at case heat management strategies and not cooler function. On really hot Summer days I could get my SSDs up that temperature, but it was mostly during gaming and because the PSU was radiating heat all over them. Do you have another way to gauge the internal case temperature? (Even just a portable thermometer would work). With a S340 the H100iv2 has to be intake. Make sure you are not using a dust filer as well, further choking the air flow.

Link to comment
Share on other sites

I am open to continuing the look at this. Your friends that run 60-68C... Is it the same stress test, with a H100i v2 (or any other 240mm cooler) and at what voltage?

 

Now that I see the Link data:

 

1) What stress test or program were you running and for how long?

 

2) I can see the end coolant temp is 41C. What was the initial value?

 

3) Your SSD and motherboard temp sensor also read near 40C. This suggests your case ambient temp may also be near the same value. If so, your coolant temp can't be lower than what is and we are looking at case heat management strategies and not cooler function. On really hot Summer days I could get my SSDs up that temperature, but it was mostly during gaming and because the PSU was radiating heat all over them. Do you have another way to gauge the internal case temperature? (Even just a portable thermometer would work). With a S340 the H100iv2 has to be intake. Make sure you are not using a dust filer as well, further choking the air flow.

 

 

1) We're both in the simulator so its not stress testing but we get around 80% CPU usage.

 

2) The initial value is 31C

 

3) Sadly I have no other way of getting internal case temperature but I had the dust filter I took it off. Is my current setup right with the S340? 184526.36ba24ba15d6f4860042535f27923503.1600.jpg

184526.68f31f9d458ed594eb892dd4d09a17f7.256c.jpg

Link to comment
Share on other sites

1. OK, so you are playing the same flight simulator. I suppose different maps could be more or less difficult to render, but we'll assume that is equal. I haven't used a flight simulator in ages and I know they are CPU intensive because of the high number of objects, however I am surprised your CPU temps are riding in the upper 70's @80% given your AIDA max at 100% is 79-82C. So, are those peak values or are the cores consistently in the 70's while playing? Next time you use it, open the stress test window from AIDA prior to loading up the simulator. It will line graph your values while you play with the window idle. You may need to configure it to show cores 1-4, VRM, or whatever else you want it to display. If these are just peak values for one instance in time, there isn't much you can do and this is increasingly the case with a lot programs. My highest temps are often when I launch the program, not when I am blasting away through a high density map.

 

2/3. 31 to 41C delta while using the simulator or any other moderate to high CPU/GPU combo load is pretty much right on the mark. 1 hour of a similar moderate CPU/High GPU game and I am +10C on the coolant delta, although with a bigger CPU and cooler. Yes, that is more than CPU stress testing alone and case heat makes up for the additional rise. You can get a rough estimate of case temp by looking at the motherboard temp sensor in AIDA. If you can compare the coolant delta (+10C) to the motherboard temp delta (before/after), it will give you a rough idea of how much is CPU waste heat and how much is general case waste heat. However, the 31C idle is what I would expect for this time of year (assuming you are not in a cold Winter) and for the S340. I am expecting a +10C number on the motherboard as well. I might be possible to shave some heat from the system by reversing the the H100iv2 fan direction to exhaust, then turning the top AF140 to intake. That would expel the CPU waste heat directly. However, you shouldn't change anything while we are trying to sort this and I suspect your friends have the same traditional set-up as you, so I don't consider this a configuration "problem".

 

Any voltage difference may be key to the discrepancy between you and your friends CPU temps. Above 1.30v+ is where the temp vs voltage curve gets steep and I would expect a noticeable difference between your 1.31v and 1.27C. I am not sure if that would amount to 10C, but you also need to make sure we are talking about peak vs peak and not average CPU temp. If you clear the AIDA graph before you launch the simulator you will be able to track the rise and fall of CPU core temps along the line graph. This may be more revealing that peak values alone.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...