What's the best non-conductive thermal paste?

[PGRacer]

For GPU & CPU, which is the best performing? I'm too clumsy for liquid metal, so non-conductive only please.

[Kaaverik]

I would say that Arctic MX-5 is the best performing non-conductive paste. It is my go-to for paste actually! Just put down a small glob and press on the heatsink. Voila!

 

If you don't want to deal with a messy paste, the IC Graphite Thermal Pad is also an option. All you need to do is cut it to size the chip (if using on a GPU) and lay it on the cpu. Press on the cooler, and Voila again! Since it is not paste, it will never dry out, and also will continuously perform at its maximum potential.

 

 

[Valentin17]

Arctic MX-2, MX-4, MX-5, Noctua NT H1/H2, Coolermaster master gelmaker etc. Buy the cheepest one because there is one or two degress diference between them.

[freeagent]

My personal faves are TFX, and SYY-157.

[tkitch]

1 minute ago, Valentin17 said:

Arctic MX-2, MX-4, MX-5, Noctua NT H1/H2, Coolermaster master gelmaker etc. Buy the cheepest one because there is one or two degress diference between them.

This.

 

 

9 minutes ago, PGRacer said:

For GPU & CPU, which is the best performing? I'm too clumsy for liquid metal, so non-conductive only please.

liquid metal is not for normal cooler installs, don't try to use it for such. 

[Chris Pratt]

Any high quality paste will perform pretty much just as well as any other. I prefer Thermal Grizzly Kryonaut, but Noctua, Artic, etc. will all work as well. It mostly comes down to personal preference.

[Valentin17]

8 minutes ago, PGRacer said:

For GPU & CPU, which is the best performing? I'm too clumsy for liquid metal, so non-conductive only please.

Before applying liquid metal you can apply some transparent nail polish to small tranzistors around the chip.

[Chris Pratt]

11 minutes ago, Kaaverik said:

If you don't want to deal with a messy paste, the IC Graphite Thermal Pad is also an option. All you need to do is cut it to size the chip (if using on a GPU) and lay it on the cpu. Press on the cooler, and Voila again! Since it is not paste, it will never dry out, and also will continuously perform at its maximum potential.

It's worth mentioning that thermal pads do not perform as well as pastes do. They mostly exist for convenience, especially for people that constantly switch out CPUs on test benches and such, like reviewers. As long as thermals are not the focus of the test, it's much easier to just throw down a thermal pad.

[PGRacer]

18 minutes ago, Valentin17 said:

Before applying liquid metal you can apply some transparent nail polish to small tranzistors around the chip.

I want to keep things valid for RMA so that's not really an option. 

[PGRacer]

20 minutes ago, Chris Pratt said:

Any high quality paste will perform pretty much just as well as any other. I prefer Thermal Grizzly Kryonaut, but Noctua, Artic, etc. will all work as well. It mostly comes down to personal preference.

Thermal Grizzly Kryonaut Extreme claims to have 14.2 W / (m K).
Almost all the others seem to have 12.5W / (m K). Is there a noticeable difference in real world temps?
I'm looking for the absolute best, if it exists. It's a custom rig that gets taken apart for cleaning twice a year, long life is not an issue.

[tkitch]

4 minutes ago, PGRacer said:

Thermal Grizzly Kryonaut Extreme claims to have 14.2 W / (m K).
Almost all the others seem to have 12.5W / (m K). Is there a noticeable difference in real world temps?
I'm looking for the absolute best, if it exists. It's a custom rig that gets taken apart for cleaning twice a year, long life is not an issue.

no, no real difference.

Differences in mounting will make more of a temp difference than the paste. 

[Chris Pratt]

7 minutes ago, PGRacer said:

Thermal Grizzly Kryonaut Extreme claims to have 14.2 W / (m K).
Almost all the others seem to have 12.5W / (m K). Is there a noticeable difference in real world temps?
I'm looking for the absolute best, if it exists. It's a custom rig that gets taken apart for cleaning twice a year, long life is not an issue.

From what I've seen it's only ever maybe a 2-3C difference, in the most extreme of cases. I haven't personally done any real comparison. I know Kryonaut performs well, and I like it. That's enough for me.