Best thermal pastes in 2021 ?

[toilet pepper]

I've done a lot changing coolers over the past few years and I've tried almost all the top pastes out there.

MX4 is cheap, has good performance.

NT-H1 consistent performance but the performance degraded on my 1070 and replacing it with MX4 after several years improved temps.

Kryonaut maybe a few degrees better than MX-4 and almost the same as NT-H1. Expensive.

ID-Cooling - it sucks.

Cooler Master Master gel - Same performance as MX4

TFX - Best paste I've tried ever. PITA to apply. Consistency is akin to snot oozing from your nose when you have the flu (including the hard parts.. I'm talking about the green stuff and not the clear stuff). I've wasted a lot of these since it is hard to apply.

 

[freeagent]

Consistency is akin to

Mine was like really soft hashish.. I should order another tube. I swap back and forth between my 5600 and 5900, it chews a lot of paste.. that's why I like SYY, nearly the same performance, much easier to spread, and its cheap like borsch, sorta.

 

[toilet pepper]

Mine was like really soft hashish..

That's it!!

Heard great things about SYY and GD900 and they come really cheap but never really wanted to risk it. I just bought a Hydronaut for my rig maintenance as it should be better than Kryonaut in the "pump-out" sense.

 

[gasolin]

PR | ARCTIC MX-5 Thermal Compound | ARCTIC

ARCTIC Press Information on MX-5 Thermal Compound | ARCTIC Service | ARCTIC Support | ARCTIC Store

www.arctic.de

 

[FreedomEclipse]

That's it!!

Heard great things about SYY and GD900 and they come really cheap but never really wanted to risk it. I just bought a Hydronaut for my rig maintenance as it should be better than Kryonaut in the "pump-out" sense.

GD900 all the way. Less scratching and pitting than kryonaut

 

[Shrek]

Yep, I got me a lifetime supply of GD900

GD900 not GD900-1

 

[toilet pepper]

Yep, I got me a lifetime supply of GD900

GD900 not GD900-1

I'd say this is where marketing and branding falls short. I'm hesitant to buy them with their packaging and it looks cheap. It simply looks like a generic thermal paste you see sold everywhere.

 

[Shrek]

The reviews say it is great.
What's The BEST Thermal Paste in 2019....? (Ft. GD900) - Bing video

Marketing drives prices up.

 

[Borc]

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.

MX-5 is the opposite of what he needs, it's a watery paste and not recommended. And KPx loses performance over time, even though this can happen on a lots of pastes on a bare die chip.

My personal recommendation is Coolermaster CryoFuze or SYY-157.

 

[Remeca]

I don't repaste often, usually only when I'm swapping CPU's. Less than once a year on average. I'll use whatever is handy, either the square of paste that comes on some HSFs, or the tube of whatever came with the cooler, which right now is whatever Noctua packages with theirs. One time I didn't have any and used diaper cream, which worked fine for a while.

 

[Shrek]

In a bind I'd also use diaper cream, but preferably the 40% zinc oxide strength.

 

[freeagent]

In a bind I'd also use diaper cream, but preferably the 40% zinc oxide strength.

Prevents fungus, chaffing, and keeps from overheating.. quite versatile..

 

[lexluthermiester]

PR | ARCTIC MX-5 Thermal Compound | ARCTIC

ARCTIC Press Information on MX-5 Thermal Compound | ARCTIC Service | ARCTIC Support | ARCTIC Store

www.arctic.de

View attachment 234032

That was one batch. New stock will be free of the problem.

In a bind I'd also use diaper cream, but preferably the 40% zinc oxide strength.

Seriously? You are never going to let the "diaper cream" thing go, are you?

 

[eidairaman1]

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.

I use a cc or spatula to spread all pastes across entire die/ihs

I've done a lot changing coolers over the past few years and I've tried almost all the top pastes out there.

MX4 is cheap, has good performance.

NT-H1 consistent performance but the performance degraded on my 1070 and replacing it with MX4 after several years improved temps.

Kryonaut maybe a few degrees better than MX-4 and almost the same as NT-H1. Expensive.

ID-Cooling - it sucks.

Cooler Master Master gel - Same performance as MX4

TFX - Best paste I've tried ever. PITA to apply. Consistency is akin to snot oozing from your nose when you have the flu (including the hard parts.. I'm talking about the green stuff and not the clear stuff). I've wasted a lot of these since it is hard to apply.

What is TFX?

 

[VeqIR]

What is TFX?

Thermalright TFX: http://thermalright.com/product-category/products/thermal-paste/thermal-paste-thermal-paste/

 

[eidairaman1]

Thermalright TFX: http://thermalright.com/product-category/products/thermal-paste/thermal-paste-thermal-paste/

I had a tube of TF4 come with the TR ARO-M14G that is on a Ry7 5800 in a Zalman Z1 Neo with 140MM fan mounted correctly (IT FIT!) I already got TG Kryonaut so i used that, glad I have spare paste to use though)

 

[Shrek]

Seriously? You are never going to let the "diaper cream" thing go, are you?

I didn't bring it up again.

One time I didn't have any and used diaper cream, which worked fine for a while.

My thing is diaper cream, yours is pump-out

 

[VeqIR]

My thing is diaper cream, yours is pump-out

And MX5.

 

[Shrek]

No offense intended; the diaper cream idea is more for amusement than anything.

But it does make me wonder how 'modern' thermal pastes are so different from pastes of old (liquid metal aside).

Forgot to add, I do like the idea of using copper based anti-seize grease, although I prefer oxides as there is nothing left to oxidize.

 

[freeagent]

No offense intended; the diaper cream idea is more for amusement than anything.

It’s better than toothpaste

 

[VeqIR]

No offense intended; the diaper cream idea is more for amusement than anything.

But it does make me wonder how 'modern' thermal pastes are so different from pastes of old (liquid metal aside).

Forgot to add, I do like the idea of using copper based anti-seize grease, although I prefer oxides as there is nothing left to oxidize.

My guess is they try to improve the temperatures while also improving longevity/durability. Which of course your diaper cream doesn't have. The longevity is the issue in terms of testing new pastes. Things like bad batches--and that also means batch variations--exist, complicating and skewing reviews. Not to mention different use cases can have different effects on longevity, so someone's experience with temperatures and longevity might differ from another person's with a totally different system and usage conditions. Even just a suboptimal mounting of a cooling plate or uneven distribution of paste can skew results enough to push a paste to the top or bottom of a comparative chart, when the differences are fractions of a degree. Also unless what's running on a test system is tightly controlled, power usage might be slightly different from one test to the next, so CPU or GPU temperatures might be up or down based on that, also skewing results. I'm seeing a fair number of contradictory "rankings" proposed by reviewers and individual users, so my personal tactic is to read as many reviews and personal experiences as possible, and then form an educated guess.

 

[Shrek]

A most excellent point; I wonder how many reviews cover longevity.

 

[VeqIR]

P.S.: if we want less biased results, then tests need to be done in some other way: on some simulated metal plates with heated elements underneath (not on CPUs) with temperature sensors and controlled cooler mounting pressure, while
1. Making multiple tests for each paste and calculating statistical averages
2. Testing all other pastes in the same controlled manner.

Such conditions are of course inaccessible to most end users, but I'm surprised there are no professional reviewers doing this kind of thing. That'd be how I'd compare pastes against each other if I had access to an engineering lab, etc. Also, we don't really care about absolute performance as much--that's a known number range already, and plenty of end users give the ballpark numbers. We mostly care about the relative performance of the pastes--and pads--with respect to each other (and longevity/durability, but that's definitely a more involved one to test). One could even turn this into a PhD project for material science research or perhaps mechanical engineering for creating the equipment and conditions needed for such a project.

Any anecdotal experience of One mounting and One set of test results is within small number statistics of uncertain precision and accuracy.

 

[Chrispy_]

A most excellent point; I wonder how many reviews cover longevity.

It's not been done for a very long time but I stopped caring because I remember sites like sharkyextreme, dansdata, and techreport (all now dead) proving that eventually they all end up the same.

The runny pastes get squeezed to such a thin layer over time that their performance is irrelevant because there's so little of it between the die/IHS and the cooler base plate.
The viscous stuff is often more conductive but gets squeezed out less, so its performance over time remains largely unchanged.

As a rule of thumb, the only longevity issues are with pastes that dry out and then exhibit movement due to poor mounting pressure. The paste drying isn't in itself a problem; microscopic valleys and troughs of the two mating surfaces are still full of the dried conductive paste. It only matters if your mounting pressure is so poor that it allows movement (which is horrendously awful in the first place) because at that point the paste is too dry to re-flow into the new required position.

TL;DR

• Long-term paste performance should be negligible because eventually the mounting pressure squeezes nearly all of the paste out of the sides anyway.
• Pastes that dry out faster are often better performers in the short term because the above action happens sooner
• Viscous pastes that dry out slower will catch up in relative performance over time.
• Dry paste is only a problem if your mounting pressure is so bad that your cooler MOVES, you're already overdue a remount and repaste at that point anyway.

 

[Borc]

A most excellent point; I wonder how many reviews cover longevity.

Basically nobody, it is very hard to test and even harder to test longevity on several pastes in the same environment. For example this test:

After 6 months longevity for one paste, imagine having 10+ pastes.