I just wanted to reply to the TIM talk.
Liquid metal does make a huge difference. More than 4c. Much more. I know, I use it. (more on that in a minute!)
But guess what... it doesn't make that big of a difference in between your cpu and water block
. I noticed like a...2c difference there (if that, it could have been due to the level of care I used applying it!) Is it better than mxwhatever or as5 or whatever shin etsu stuff you can buy, or various other things? Yes. How much better? Couple degrees maybe. Maybe. If it's more you are probably doing something wrong applying your other stuff. Or your stuff isn't flat.
It works fine on copper. It works fine on nickel. Don't put it on aluminum. At least, if it's the same liquid metal product I've used. The contents are not a secret (the main ingredient starts with a "G" in most of them). You can put it on an aluminum can to test it!
So where do those huge numbers people find (including myself, I noticed a massive almost unbelievable difference) come from? Well, it's off topic to this thread.
It's on de-lidded ivy and haswell where there is a massive gap between the heat spreader and die. Even sometimes after you clean the gunk off. It works *great* there because there is such a huge (in comparison) gap between the die and the heat spreader. I'd imagine if you cleaned the rim of the heat spreader off real well to make it fit super nicely the difference would be smaller. But I didn't want to take sandpaper to my heat spreader's rim.
If there is some sort of massive gap(s) between your heat spreader and your block you are probably doing something wrong lol.
When I put it in between the die and spreader on my 4770k I noticed a huge (I'm not going to say for sure as it was so long ago but it was >10c, easily) drop. When I replaced the mx4 in between the water block and heat spreader temps went down like 2c. I wouldn't use it in between the block and spreader again as who knows how it'll clean off after years of use. And that difference may have been how careful I was in applying it--I cannot say for sure, there's no way for me to really judge how well I applied it!
So I don't really think this should apply to AMD users. I didn't use it on my 8320 pc (just used the stock stuff that came on the h100 I used in that, but I havne't fired that PC up in a long time).