Best thermal pastes in 2021 ?

[jsalpha2]

Arctic Silver Ceramique Been using it for years. If it isn't broke...

 

[Hankieroseman]

I bought Corsair XTM50 and going to give it a try with Ryzen 7 3800X and the stock cooler.

Not very fond of the XTM50. It's rather thick and doesn't spread like Arctic Silver. It's odd with a template and the wide spreader. I wouldn't use it again. Ryzen 7 idles at 44 Celsius.

 

[Falkentyne]

Thermalright TFX is one of the best pastes out there and definitely the most durable (just let it cure first, before putting it under >75C temps, for about a week).

 

[Keullo-e]

I made a general thread about pastes already https://www.techpowerup.com/forums/threads/the-official-thermal-paste-thread.278074/

But I like MX-4 the most and I recommend that.

 

[abactuon]

best thermal pastes in 2021

The same thermal pastes as 2020, 2019,2018 ...

 

[Falkentyne]

TFX is one of the most durable pastes now and is the best for laptops. If you apply it correctly (do not spread it), it's also one of the top performers for desktops and video cards after it cures for about a week!

 

[doyll]

I've using both TFX and TF8. TFX is thicker but a few minutes in hot water makes is easy to use. TFX is rated 14.3 W/mk to TF8 being raed 13.8 W/mk, so both are among best on market now, unless you go liquid meta / gallium based TIMs which are much harder to work with. They can damage things if not careful.

 

[Kaliphtan]

I made a general thread about pastes already https://www.techpowerup.com/forums/threads/the-official-thermal-paste-thread.278074/

But I like MX-4 the most and I recommend that.

You mean a day after my thread

I've using both TFX and TF8. TFX is thicker but a few minutes in hot water makes is easy to use. TFX is rated 14.3 W/mk to TF8 being raed 13.8 W/mk, so both are among best on market now, unless you go liquid meta / gallium based TIMs which are much harder to work with. They can damage things if not careful.

Be careful, MK rating and stuff like that usually don't mean nothing much, check the benchmarks on Youtube or feedback, by the way I got myself an Arctic MX5 last september, it's working nicely, same performance as the Noctua H2.

 

[VeqIR]

Would you recommend TFX, SYY-157, or NT-H2 on a GPU? It’s time to repaste an old card. I already own the second two.

 

[lexluthermiester]

Would you recommend TFX, SYY-157, or NT-H2 on a GPU? It’s time to repaste an old card. I already own the second two.

MX-5 or KPx.

 

[VeqIR]

MX-5 or KPx.

MX-5 seems to be around on par with my NT-H2 that I already have, but NT-H2 is much easier to use. I was thinking about something thicker/denser for low pressure mounts, like for GPU cooling. The things that scared me away from MX-5 in general were the highly "sticky" behavior and reviews on Amazon describing how MX-5 can creep around the edges of the CPU integrated heatsink and actually travel underneath onto motherboard pins... I've never ever had that happen with a paste and don't want there to be a first personal experience of such an event.

Based on one recent large review, NT-H2 does fine on low pressure mounts too, so I'll probably just use that.

Best Thermal Paste for CPUs 2024: 90 Pastes Tested and Ranked

The right thermal paste will improve performance.

www.tomshardware.com

One thing I regret is that these reviews often don’t give sufficient curing time to pastes that need it, so results are not completely representative of longer-term performance.

 

[lexluthermiester]

but NT-H2 is much easier to use.

However NT-H2 dries out and loses performance over time whereas MX-5 and KPx do not.

I was thinking about something thicker/denser for low pressure mounts, like for GPU cooling.

Then you want MX-5 or KPx as they are perfectly suited for such use.

The things that scared me away from MX-5 in general were the highly "sticky" behavior

If you'd like to see the testing I did look here;

[EOL] Arctic MX-5 is here!!Tests incoming! Completed. Now its MX-6 testing time!

Oh I'll pass on that. That's specifically for TFX. (I got the best results on my 3090 FE (most stable core temp deltas) applying it like that. SYY-157 still needs full coverage but is easier to get on. Anyone here use SYY-157 yet?

www.techpowerup.com

and reviews on Amazon describing how MX-5 can creep around the edges of the CPU integrated heatsink and actually travel underneath onto motherboard pins...

Total nonsense.

I've never ever had that happen with a paste and don't want there to be a first personal experience of such an event.

Neither have I and you're not going to.

One thing I regret is that these reviews often don’t give sufficient curing time to pastes that need it

That's because there is no curing time. Arctic Silver 5 was one of the only TIMs that needed curing time, and nothing modern needs it at all.

so results are not completely representative of longer-term performance.

Sure they are.

 

[VeqIR]

Do you have a link to discussions of NT-H2 losing performance? (in non-laptops). And the comparison you linked to compared against NT-H1.

 

[eidairaman1]

However NT-H2 dries out and loses performance over time whereas MX-5 and KPx do not.

Then you want MX-5 or KPx as they are perfectly suited for such use.

If you'd like to see the testing I did look here;

[EOL] Arctic MX-5 is here!!Tests incoming! Completed. Now its MX-6 testing time!

Oh I'll pass on that. That's specifically for TFX. (I got the best results on my 3090 FE (most stable core temp deltas) applying it like that. SYY-157 still needs full coverage but is easier to get on. Anyone here use SYY-157 yet?

www.techpowerup.com

Total nonsense.

Neither have I and you're not going to.

That's because there is no curing time. Arctic Silver 5 was one of the only TIMs that needed curing time, and nothing modern needs it at all.

Sure they are.

I'd still use AS5/Ceramique myself along side A MX series, CM MGM, Gelid, TG Kryonaut. I will not use conductonaut or noctua, or liquid metal/diamond based pastes.

Haven't used hydronaut.

 

[VeqIR]

I will not use conductonaut or noctua, or liquid metal/diamond based pastes.

Out of curiosity, why not Noctua pastes?

 

[eidairaman1]

Overrated.

 

[VeqIR]

Overrated.

So just the prejudice? No personal experience? There are plenty of comparative reviews with their pastes.

 

[freeagent]

I didn’t like Noctua paste either. It is overrated, along with some of their other products.

 

[lexluthermiester]

Do you have a link to discussions of NT-H2 losing performance? (in non-laptops).

Those discussions have happened here at TPU in places. Can't remember exactly where. It's also happened in other forums.

And the comparison you linked to compared against NT-H1.

[EOL] Arctic MX-5 is here!!Tests incoming! Completed. Now its MX-6 testing time!

Interesting results with MX-5, meh... personally I'll stick with thermal grizzly range of TIMs. :)

www.techpowerup.com

This was a similar discussion. Toms Hardware had a good comparison of performance.

 

[Vario]

I didn’t like Noctua paste either. It is overrated, along with some of their other products.

I had a bunch of tubes of NTH1, I ended up giving them all away. The NTH1 would always yield great results immediately but after a few weeks worsened, unmounting the heatsink and it was apparent it thins out.

I like AS5 because it is very stable. Right now I have a Radeon 7850 2GB graphics card running 9 years on the same AS5 mount with the same temperatures as when installed. I use a plastic bag on my finger tip and spread the AS5 thin and uniform.

 

[freeagent]

I had a bunch of tubes of NTH1, I ended up giving them all away. The NTH1 would always yield great results immediately but after a few weeks worsened, unmounting the heatsink and it was apparent it thins out.

I like AS5 because it is very stable. Right now I have a Radeon 7850 2GB graphics card running 9 years on the same AS5 mount with the same temperatures as when installed. I use a plastic bag on my finger tip and spread the AS5 thin and uniform.

That is exactly what I saw too. AS5 definitely beats NTH1.

 

[FreedomEclipse]

Im using HT-H2 on just about everything -- all my laptops, GPUs and CPUs. been running it for 2-3 years and never had a drop in performance.

In any case, I tend to steer clear from diamond particle based pastes because i dont want any pitting or scratching on mating surfaces. Chasing that 2-3'c aint worth the risk to me and my 8600k and 1080Ti are pretty much maxed. My 8600k in particular wont go any faster.

 

[freeagent]

Glad they fixed it with H2. H1 was unbecoming of the Noctua name.

 

[Vario]

That is exactly what I saw too. AS5 definitely beats NTH1.

Here is my idle (19C room ambient) and load (P95 blend) with AS5 that was put on my i5 8600K a year ago. I don't see a reason to use other pastes. This particular AS5 syringe was bought in the mid 2000s as well, so it may be 15 years old.

 

[freeagent]

I don't see a reason to use other pastes

For me it was when I moved to 3rd gen Intel. I tried out the paste that came with my Le Grand Macho RT, the paste was CFIII and it held up pretty well against AS5 for the short term. Then I got curious and tried some others.. I still think of AS5 with fondness, but I think they should play with their formula a little bit, or maybe reformulate something that can better deal with the unique heat load of modern CPU’s. The mid 2000s was a good time for AS5, I’ve remounted so many times over the years that I have got my process down pat lol.