Jump to content
Corsair Community

h100i Gtx temps...


Recommended Posts

Every few years I try to add a few new parts to my build in an attempt to mitigate the the big hit at one time in buying a new pc. Coming from an older antec 300 case I decided to purchase a Enthoo Pro full tower and while I was at it get an h100i gtx to replace a hyper 212 evo. The move from cases went smoothly but not sure Im getting the performance that I should out of the h100i gtx. Maybe my expectations are too high for an AIO. Let me know what you think.


Running a z87 platform with an i5 4670k clocked at 4.2ghz at 1.2 volts in prime 95 small FFTs with maxed out fan speeds on h100i gtx Im getting temps as high as 84c on my hottest core. As soon as the test starts Im in the high 70s and by a min into the test my hottest core is over 80. If I let test run a few mins the pump never feels really especially warm or hot to the touch. In fact the entire unit pump, hoses, or radiator never feel like they are getting really hot.


I have attempted to reseat the block/pump and while the standoffs are not tight and have play in them with no pump once the pump is seated and the thumbscrews are then tightened everything is good and snug with no movement.


Any Ideas? the Link software seems to be working ok and giving me readings like everything is operating fine but those temps seem quite high.

Link to comment
Share on other sites

Are you saying you were getting similar results with the 212? Running small FFT Prime on Haswell is just asking for ego melting numbers. You would be better off using another test program, unless you are very familiar with how to set up custom runs and have meaningful 212 data to compare.


Water cooling works a little bit differently to air. Your CPU temperature is going to be [water temperature (H100i GTX temp) + voltage induced CPU temperature] = Final CPU temperature. There isn't anything you can do about the voltage induced temperature increase, other than lower the voltage. The other thing would be to delid, which you have already done. Where you have some more control is over the water temperature. This will be your case ambient temperature to start, plus whatever heat to you add (CPU, GPU, other hardware), less what the fans/radiator are able to remove. The fans will not cool the CPU directly. Instead their job is to expedite the removal of heat from the water as it passes through the radiator. The water will transport heat from the contact plate as fast it can flow. This is why everything felt relatively cool to the touch. The system is working and it is keeping water temperature down as much as it can. However, none of that can reduce the voltage induced heat on the chip itself. Your air tower can't do this either, and you also don't have any control over the heat transmission through the metal piping.


One way to verify this is to look at the H100 GTX Cooler Temp (or H100 GTX Temp) in the LINK program. This is your water temperature. It should be slow to change, both up and down. The fans on the GTX will base their speed on this value, rather than CPU temperature, since their job is to remove heat from the water. Try another run with AIDA, OCCT, Intel XTU (anything else really), and watch the water temperature value. This will give you a better idea if you system is working efficiently. You don't need to reseat the block. If it was off, you would hit the boiling point a fraction of second after initiating Prime and your idle temperatures would be rather high as well.

Link to comment
Share on other sites

It does sound high because I have the same chip z87 pro board in my kids computer oc'd at 4.3 with 1.25 volts and i hit about 74 with intel burn test. Im using the h80i inside the case at the back as exhaust. I did nothing to the cpu itself, its stock.


That's not the same test as Prime 95 Small FFT and we have no idea which version was used.

Link to comment
Share on other sites

Thanks for all the info C-attack. Its been about 2 years since I have messed with overclocking and setting this rig up with the Evo and never bothered benching it or testing it again in Prime to see the numbers vs the h100i gtx. Running Prime 95 blend temps are much lower than the 85 or so I was getting with the Small FFT's.


Using Realbench h.264 video encoding had me up to 65 on the highest core and 60 on the lowest core at 4.3ghz at 1.95 vcore. With all that being said it seems the h100i gtx is working just fine and I just didn't realize how brutal prime 95 small FFTs are on haswell in particular.


Being a real noob who has only dabbled a bit in overclocking an AIO's as well I have 2 other questions if you don't mind me picking your brain.


1. While using none cpu intensive programs is it best to keep the pump in quiet mode to reduce the RPM's in an attempt for longer pump longevity and only switch it to performance when gaming?


2. Is there a given h100i gtx water temp I should be concerned about ever exceeding. So far Ive never really seen the water temp over low to maybe mid 40's.

Link to comment
Share on other sites

1. This seems to be a popular question today. Short answer: it doesn't matter. Use it how you need or prefer.


However the longevity of the pump is measured (hours of use or revolutions), the product has to have a predicted lifespan of close to 5 years, otherwise both the OEM and Corsair would suffer heavily on the warranty. Assuming the pump does have some sort if linear degradation based on the speed (and that is an assumption), would you prefer to nurse and baby the unit for 5 years? Or use it as you need and know that you can have a replacement in short order during those 5 years? By the time warranty is up, you are likely looking for something different anyway.


When you read about someone's pump failing in the first year or so, it's not because they "overused it". Some piece in the mechanical parts was not up to specification or an imperfection in assembly. If your pump fails at 14 months, it is not because you used performance mode the whole time. Something else was wrong. And you are still under warranty.


Use the pump the way you need to. I might put it in Quiet mode while on the desktop simply because I spend so much time working/reading from this PC. However, there is another side to this. Running a faster pump speed may allow you to run even lower fan speeds because the water can carry more heat to the fans in a given amount of time. This isn't very relevant on the desktop, but can be an alternative on moderate to high loads.



2. At 100 Celsius I would quickly leave the room. More seriously, most people should be under 50C regardless of room temperature. However, that is the big factor. Water temperature = (case ambient temperature + CPU waste heat) - the the amount of heat the fans are able to remove through the radiator. So, my Winter time water temp when I keep the house at 16C will be decidedly different than my Summertime room temp of 27C. GPU waste heat is part of the case ambient temperature. Someone with 980 Ti x 2 will have a lot more case heat and higher starting water temperature than someone with a smaller GPU. So, there is a lot of variance as to what your water temperature should be, based on both room temp and hardware. The average user in a temperature controlled room probably runs between 25-27C on cold boot and 40C under sustained load. The longer you run, the more likely the water temperature will rise. This also means most people need to make their own custom water temp vs fan speed curve. If you have heavy GPU's or warm room, you are likely to have a lot of fan speed that may not really be helping you.


I don't see anything abnormal with your 40C result under load, but compare it to future results and note the room/case temperature before and after to see whether the differences come from additional ambient heat or less heat is being removed through the fans (slower speeds).

Link to comment
Share on other sites


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

  • Create New...