Originally Posted by TwoCables
Hmm. I'm not sure, but probably only a few degrees. Nothing to worry about. It's not about the temps, it's about making sure to test the entire CPU instead of just a part of it. So, run the Blend test with 28.5.
Wouldn't surprise me for a second if it was 30c hotter
I don't think that orndorf will want to run it
This is Haswell, where 100% CPU load with a video encoder can draw 70w from the CPU while 100% load with Linpack getting 210gflops @4ghz can draw 130w. He's nowhere even remotely close to max temperatures now - if that old prime is anywhere near as cold as Realbench, then he can gain about 50c just from switching stress test.
do you know of any custom water loop kits that come with a radiator the same thickness of the corsair h100i ?
As for this.. A water setup with a thin 240mm radiator like the h100i would outperform it, but not by a large leap. Water is only good when you can use more radiators - like for example in my case, i would consider running 2x 240mm and 1x 360mm. That's 7x 120mm fan width instead of the 2x that you'd have using a radiator the size of h100i's - i could also potentially use thicker radiators for some slots
Water is also harder to do in a case with multiple GPU's. Either you do a full custom loop including the GPU's, at which point you need a LOT of radiator area or your CPU temperatures will be hurt by the 500w of heat that your GPU's are adding to the water - or you don't include the GPU's, and you are starving them of airflow by having radiators on your fan mounts blocking airflow. They're either exhausting hot air through the case through them, or blocking the GPU's cold air - neither works well. You need certain case setups for 2x high end GPU + a CPU water setup to work well.Edited by Cyro999 - 5/16/14 at 7:43am