Liquid Metal temperature drop

I have an i7 4790k that has been over clocked to 4.6 or 4.7 at 1.25v for over a year. Max temps were 75C (noctua nh-d15) in every test other than Prime 95 28.10 where I would hit 95C max. Well today I finally took the heat spreader off the CPU and used Cool Laboratory's Liquid Ultra below and above the heat spreader. I have read what others were saying about their results and I thought they were being optimistic on their results quoted. Well after doing mine I found my high temps in prime 95 are now 78C on 28.10.

That is a massive drop in temps by just changing the TIM. I was thinking of switching to water cooling but I figured I would try the TIM change and see. Results of 17C drop made me rethink water cooling. How many others in here have done this?
 
Be careful with that liquid metal: if it tries to terminate you and Arnold isn't around you'll need to find a steel plant to destroy it in on your own!

In seriousness, I've used different TIM (and noticed modest differences) though haven't gone to the length of popping off the CPU heat spreader. I'm a bit surprised such a big difference was found without a new cooler.
 

CompuTronix

Intel Master
Moderator
I've been delidding and using Liquid Ultra ever since Intel switched from Indium solder on Sandy Bridge to TIM on Ivy Bridge. My 3770K and 4770K were both bare-die mounted air cooling, but as the substrate is much thinner on my 6700K and 7700K, it's necessary to use the heat spreader. All were heavily overclocked.

Hottest Core on my 7700K is 73C at 22C ambient with P95 v26.6 Small FFT's. Across 5 processor generations I've been running the same Cooler Master TPC 812 with dual fans and see no need for liquid cooling, or the inevitable pump failures that we see so frequently on our Forums.

CT :sol:
 


On the 26.6 version of Prime I hit a max of 60C.