On one hand, I have a virgin Q6600 G0 processor, begging to be overclocked.   On the other, I have various thermal compounds that have accumulated over the last few weeks.  Among them, some familiar names, others brand new.  Which should I use?  Why not test them all and find out?

Baseline:

     Of course, to test anything, we need a control for our experiment.  So why not use the stock paste?  Since I will be using the stock cooler, I might as well test the stock paste as well, which is categorized as TC-1996.  So, to get a baseline, I am using the new Quad Core aware version of Prime95, version 25.4.  To monitor temps, I am using SpeedFan, logged over a 20 minute interval; 5 minutes idle, and 15 minute load.  All the results of all testing will be combined at the end of the article.

For the purposes of this test, the CPU will remain at stock settings in the BIOS.  This particular G0 runs at 1.225 volts.  For the heatsink, all tests are done with the stock heatsink at 100% fan speed.  Case airflow is kept to a minimum. 

Arctic Silver 5:

     Arctic Silver 5 has been the thermal compound of choice for many enthusiasts for a very long time, so it only makes sense that we throw it in the roundup.   According to the instructions on the Arctic Silver web site, correct application is a straight line across the two cores.  Available in 3.5g and 12g sizes, AS5 is the old favorite.

Ceramique:

     Ceramique is the new contender from Arctic Silver, and has already won the hearts of many.   The application of the Ceramique is the same as that for the AS5.  It is generally a bit more "gooey" than AS5, but still easy to apply.  Has the younger sibling unseated the king?  Read on to find out. 

AI Technology's Cool Silver

     The first newcomer on the list is AI Technology's Cool Silver.   According to the AI Technology website, it bests the "Industry Standard" by a few degrees. I wonder who that could mean?  The paste was a bit easier to apply than the AS5 and the Ceramique.

     Our final entry is the IC Diamond 7 Carat compound.  The IC Diamond website makes a direct comparison to AS5, boasting a 5 degree fahrenheit difference.  According to the instruction, we are supposed to spread out a "thin layer" on the CPU and let set for 10 minutes, so that it can bond the ground diamond crystal to it.  Unfortunately, spreading the IC Diamond is like spreading dried out Play-Doh.  I ended up scraping off the compound and applying it in a straight line.  Now, how about some results?

     From this graph it is a bit difficult to see exactly what we can conclude.  What we can see for sure is that the old standby AS5 wins by a large margin in terms of idle temps.  Let's see if we can make the load results a bit easier to see.

     What I have done here is isolated the last 300 seconds of the log, and magnified.  While there isn't a clear winner, it would appear that the most effective pastes are the Cool Silver, the Ceramique, and surprisingly, the stock grease from the Intel retail heatsink.  The AS5 and the IC Diamond both performed a few degrees worse.  Now for a bit of a caveat; I did not let the heatsink pastes settle for the 200 hours that the Arctic Silver products recommend, mainly for the interest of time.  The IC Diamond blatantly says on their website that no settling time is needed for their product, so that is a non-issue for them, and no mention of it on the Cool Silver site. 

Conclusion:

     So what have we learned, girls and boys?  For one, the stock thermal paste isn't nearly as bad as it used to be.   In fact, you could almost argue that it won this particular competition.  The Ceramique and the Cool Silver both performed admirably, while the IC Diamond has not lived up to their own hype, and the Arctic Silver 5 has been dethroned.   Perhaps the AS5 could redeem itself after re-settling, however, past history indicates that is only good for a couple of degrees.  In reality, however, any of the above pastes will do the trick in a pinch, so always make sure you have some handy!