CPU Heat Spreader Polishing

I recently read somewhere that prior to "water" cooling a CPU that has a large heat spreader, like all the newer CPU's do, that the top of the spreader should be polished to provide maximum contact with the water block. First of all; has anyone else read anything similar??? Secondly; The spreaders seem to be made of very thin material, consequently, what would be the proper approach without risking the integrity of the CPU? 1000 grit wet and dry sandpaper, jewelers rouge, steel wool, ..... ????:hyper

 

I lap all my CPU & Chipset HS, sometimes the CPU lid
Here's how to link http://www.overclockersclub.com/guides/heatsinklappingguide.php

I go no higher than 1200 grit, wet sanded
IMO anything beyond 1200 is pointless (I'd bet you wont see diff in temps between 800 & 1200, it just makes photo's of your HS prettier)

I dont think you have to worry about sanding thru a HS and if you do it would be time to buy a new one "LOL"

 

ASUS, Thanks for the information, it sounds as though you may have had some experience in liquid cooling. Therefore, I'm looking for any help I can get, and, the answer to one of those life long questions; Does using a liquid cooling set up provide a measureable and/or meaningful differential in temperature to go through all the trouble? I've wanted to do this for a long enough time that I've reached the age where I've become practical. I'm planning on cooling a FX-57 and an 8800 GTS which both put out a lot of heat, but, if I can get equal specs with an air rig, why use the liquid approach?

 

Liquid cooling expert, I'm not
I've been thinking about Liquid for two yrs

I just built my first liquid cooled system about a month ago, The whole system is made from Used but Quality parts
For the moment I'm only cooling my CPU, strictly for OCing purposes
So far I've run two socket 939 chips
I'm seeing about 10 degree difference between idle & load temps
Compared to the same chips cooled by air, I'm seeing approximately a 6 degree improvement of Load temps on liquid, while running more cpu voltage

 

ASUS, Which BTW is one of my favorite component Mfg's, are your results in Celsius??? If so that tells me that it's well worth my time and effort. I have built air-cooled systems that had fabulous results, however, it also shook the house and years later I still have a buzzing in my ears. Based on your results, it appears that I should take the time based on the CPU/GPU's I'm going to be using. If anyone else has similar or opposing results please chime in!!!!!!