The League of Thermal Paste

mayebaza · 16 May 2011 at 22:43

According to Hardware Secrets’ latest test results there is only the five degrees between the best and the worst thermal paste. I think I will be sticking with good old AS5 until the little tube runs out. I have an unopened tube of Noctua’s NT-H1 which came with DH14; that will be going in the bin. In the future, I will try Tuniq’s TX-3 but for the extra 1 degree, it can wait unless anybody out there can tell me anything different besides curing time and non-conductivity.

http://www.hardwaresecrets.com/article/Thermal-Compound-Roundup-May-2011/1249/1

-----------------------------------------------------------------
Gigabyte GA-MA790FX-UD5P, 955 BE(C2)@ 3.92GHz 1.52V, G. Skill RipsawsX 8GB DDR3-1600,
Geil Value 4GB DDR3-1600, ATI 5850 1GB(Crossfire), CM90 II Advanced, 1TB Western Digital Caviar Black,
4 HDDs, XFX 850W BE, Asus Xonar Essence STX, BlackGold HD DVB-T2, Dual Digital DVB-C.

Phenomenologica · 16 May 2011 at 22:51

Since both room temperature and core temperature readings have 1 oC resolution, we adopted a 2 oC error margin, meaning temperature differences below 2 oC are considered irrelevant.

So basically they just said that their results are irrelevant. Taking 35 as the modal value, only two compounds fall outside their margin of error, plus the test with no compound. An excellent and informative test, then. :rolleyes:

C64 · 16 May 2011 at 22:53

well I know it's wrong already when I see arctic silver 5 above mx4

setter · 16 May 2011 at 22:57

I found mx3 better than as5, current cpu is coated with the prolimarech stuff, on a par with mx3 from my own usage experience.

panyan · 16 May 2011 at 23:10

mayebaza said:
I have an unopened tube of Noctua’s NT-H1 which came with DH14; that will be going in the bin

*cheeky voice* can i have it please?

mayebaza · 16 May 2011 at 23:14

Hmm could it be some thermal paste favour Intel over AMD and vice versa?

Phenomenologica · 16 May 2011 at 23:19

mayebaza said:
Hmm could it be some thermal paste favour Intel over AMD and vice versa?

I heard arctic silver 5 talking about Intel behind his back earlier, so you could be right. Or, thermal compound could just be a material designed to conduct heat, and thus only vary in performance depending on temperature and consistency of application.

I believe there is a slight tendency for Intel chips to run hotter than AMD ones so if there was a paste that was more efficient at higher temperatures it's possible it would 'favour' intel, but I imagine any that perform better at the high end will do likewise at the low end.

Operation · 16 May 2011 at 23:24

Phenomenologica said:
So basically they just said that their results are irrelevant. Taking 35 as the modal value, only two compounds fall outside their margin of error, plus the test with no compound. An excellent and informative test, then.

Yep. I'd suggest application is far more important than brand, and responsible for pretty much any perceived variation.

Phenomenologica · 16 May 2011 at 23:39

The whole test seems dubious. The error margin is large, they didn't repeat the tests (Including reapplying) several times and take an average, they probably didn't take into account curing times for any that need them, they don't mention how long they ran the prime test for (That I can see)...It's hardly reliable data!

gropingmantis · 17 May 2011 at 00:23

No IC Diamond?

JeffyB · 17 May 2011 at 00:41

Skinnee also noted that the prolimatech pk1 was a great performer and was only bettered by indigo xtreme.

http://skinneelabs.com/2011-tim-results/

Mozzila · 17 May 2011 at 01:25

IC Diamond should win on that chart and it's not just about which compound it is, it's how you apply it.

23JMZ07 · 17 May 2011 at 01:31

IC Diamond is way over hyped...

Mozzila · 17 May 2011 at 03:56

I read good stuff about it, i will be buying some anyways.

Does anyone know what the actual best thermal paste is with the best method of applying?

From this date ty

DamoTourettes · 17 May 2011 at 18:33

Indigo Xtreme - I've found it's *at least* 2-3 degrees cooler than other thermal pastes, however applying it makes you absolutely brick it when your cooler hits 100 degrees... haha...

Krony · 17 May 2011 at 18:57

MX-4 on my GTX 580 (Koolance block) 28c idle and 46c load.

trisp · 17 May 2011 at 18:58

gropingmantis said:
No IC Diamond?

Were is the diamonds?

Rroff · 17 May 2011 at 19:01

C64 said:
well I know it's wrong already when I see arctic silver 5 above mx4

With proper application I've never seen much difference between AS5 and MX3/4, MX3/4 is just a bit more forgiving with poor(er) application i.e. stick a big blob on the CPU and smear it around and MX4 will win by a mile, but thats because the technique is poor.

Randall8686 · 17 May 2011 at 19:33

Rroff said:
With proper application I've never seen much difference between AS5 and MX3/4, MX3/4 is just a bit more forgiving with poor(er) application i.e. stick a big blob on the CPU and smear it around and MX4 will win by a mile, but thats because the technique is poor.

+1 its more application than brand apart from real cheap stuff