• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Best thermal paste

Associate
Joined
3 Feb 2017
Posts
1,480
Location
Mondas
MX4 was not great for a Vega64 repaste. Although to be fair that GPU does get silly hot. I found the noctua paste to be very good.
 
Man of Honour
Joined
13 Oct 2006
Posts
91,640
MX4 was not great for a Vega64 repaste. Although to be fair that GPU does get silly hot. I found the noctua paste to be very good.

For large surface areas like GPUs you need a different type of paste really - ideally something designed for manually spreading.

IIRC when I was testing awhile back MX-2 was actually one of the better ones for GPUs out of what I had available like AS5 and MX-4, etc.
 
Last edited:
Man of Honour
Joined
13 Oct 2006
Posts
91,640
Any thoughts on this, or am I overthinking it (which would be rare for me actually lol)?

Personally I've never had any problems with longevity with anything other than generic pastes which come bundled with some hardware and earlier batches of Thermal Grizzly stuff (though supposedly that has been fixed). AS5, MX-4, etc. has held up for like a decade plus in some applications for me when applied properly.
 
Soldato
Joined
25 Oct 2010
Posts
5,646
Personally I've never had any problems with longevity with anything other than generic pastes which come bundled with some hardware and earlier batches of Thermal Grizzly stuff (though supposedly that has been fixed). AS5, MX-4, etc. has held up for like a decade plus in some applications for me when applied properly.

Likewise. I know people that religiously tear down their PC's to repaste/clean every few months. Outside of perhaps enjoying the process, it's an utterly pointless exercise with 95% of thermal pastes/use cases, at worst they're actually increasing the chances of something going wrong with needless faff.

Blast your rig with some air once in awhile and monitor temps, it's rare that paste longevity will be a problem.
 
Last edited:
Man of Honour
Joined
13 Oct 2006
Posts
91,640
I know people that religiously tear down their PC's to repaste/clean every few months. Outside of perhaps enjoying the process, it's an utterly pointless exercise with 95% of thermal pastes, at worst they're actually increasing the chances of something going wrong with needless faff.

Sometimes that is due to people persisting in using incorrect/sub-standard application methods - even when it might not always impact performance it can impact lifespan - incorrect hand spreading of AS5 for example can result in it needing to be reapplied every couple of years or so.

I could look it up but IIRC correctly applied MX-4 should on average last at least 7 years or something with no performance degradation (EDIT: 8 years shelf life, 7 years applied).
 
Last edited:
Soldato
Joined
25 Oct 2010
Posts
5,646
Sometimes that is due to people persisting in using incorrect/sub-standard application methods - even when it might not always impact performance it can impact lifespan - incorrect hand spreading of AS5 for example can result in it needing to be reapplied every couple of years or so.

I could look it up but IIRC correctly applied MX-4 should on average last at least 7 years or something with no performance degradation.

I have MX-4 on a few things running on close to 5 years with zero issues, it's hard to go wrong with in my experience.

I think a lot of people are of the mindset that "general maintenance = good" but take it a step too far. I've lost count of the number of people that come to forums (this one included) to complain about things going wrong after a repaste, and when questioned they usually had no reason to have done one other than the idea it was necessary just in case.
 
Soldato
Joined
15 Sep 2008
Posts
2,589
I suffered thermal paste pump out on my 4090 with MX6 after about a month or so.

embed


I didn't know what this was at the time so repasted it with MX6 again :D. The pump out may occur again...

If it does I'll use the PTM7950 stuff that many recommend for direct die applications.
 
Man of Honour
Joined
13 Oct 2006
Posts
91,640
Yeah for surface areas that size you want a thin spread layer of a more viscous paste, a blob of something like MX-4/6 isn't ideal on something like a 4090.

Intel CPUs especially are a bit different as you usually have a rectangular package under the IHS which doesn't take up the whole area and the surface is often less than flat.
 
Last edited:
Soldato
Joined
28 May 2007
Posts
18,548
I suffered thermal paste pump out on my 4090 with MX6 after about a month or so.

embed


I didn't know what this was at the time so repasted it with MX6 again :D. The pump out may occur again...

If it does I'll use the PTM7950 stuff that many recommend for direct die applications.

The contact looks pretty poor. Some kind of pad or a better fitting cooler might be the way to go.
 
Associate
Joined
20 Aug 2020
Posts
2,059
Location
South Wales
Been using PTM 7950 on the Steam Deck for a long time and it's been great, had enough left over to try it on the 7950X3D and it's seems to be working very well. Pain to apply a larger size even when it's been in the fridge for a few hours but so much easier than spreading or hoping the pea method does a nice coverage as it's always a perfectly even amount every time.

Other than that I've been using Prolimatech PK3 for years on my CPUs and GPUs and it seems to last for ages, easy to spread and doesn't dry out quickly.
 
Last edited:
Man of Honour
Joined
13 Oct 2006
Posts
91,640
does a nice coverage as it's always a perfectly even amount every time.

Unless the surfaces are very well finished/lapped you want a paste which will as best possible try and optimally fill the space - having a perfect even coverage with a less than perfect finish isn't optimal. Which is why it makes me facepalm when people persist with hand spreading AS5 in situations where doing so is suboptimal.
 
Soldato
Joined
28 May 2007
Posts
18,548
Unless the surfaces are very well finished/lapped you want a paste which will as best possible try and optimally fill the space - having a perfect even coverage with a less than perfect finish isn't optimal. Which is why it makes me facepalm when people persist with hand spreading AS5 in situations where doing so is suboptimal.

Just need to be flat. After around 800 grit the finish irrelevant. Go to 1000 if you must, but it will make zero difference.
 
Man of Honour
Joined
13 Oct 2006
Posts
91,640
Just need to be flat. After around 800 grit the finish irrelevant. Go to 1000 if you must, but it will make zero difference.

Flat is the big one but once you get to a certain level of finish you hardly need much thermal paste at all to get optimal performance, as per the Arctic Silver documentation I linked to before.
 
Soldato
Joined
28 May 2007
Posts
18,548
Flat is the big one but once you get to a certain level of finish you hardly need much thermal paste at all to get optimal performance, as per the Arctic Silver documentation I linked to before.

Trust me, I’m professional. At 800 grit you are already sub micron and particles within the TIM become an issue. The finish is completely irrelevant past around that point.

Temps will get worse the more you polish. Flat and approaching sub micron scratches is what you want to achieve.
 
Man of Honour
Joined
13 Oct 2006
Posts
91,640
Trust me, I’m professional. At 800 grit you are already sub micron and particles within the TIM become an issue. The finish is completely irrelevant past around that point.

Temps will get worse the more you polish. Flat and approaching sub micron scratches is what you want to achieve.

You'll need special polishing pastes or similar to get a finish where the particle sizes in a paste like AS5 are an issue, even 1000 grit won't get close to that.

(Arctic Silver 5 is rated down to 2000 grit mirror finish before performance is impacted by particle size).
 
Last edited:
Soldato
Joined
28 May 2007
Posts
18,548
You'll need special polishing pastes or similar to get a finish where the particle sizes in a paste like AS5 are an issue, even 1000 grit won't get close to that.

(Arctic Silver 5 is rated down to 2000 grit mirror finish before performance is impacted by particle size).

Go for it.

For everyone else, no point lapping past 1000 grit. Flat and lapping upto 800 grit in stages is all you need. 1000 if you must.
 
Soldato
Joined
30 Jul 2005
Posts
19,504
Location
Midlands
Liquid metal has far too many potential downsides really unless min-maxing a setup, phase change materials can be tricky to get the best results from - although you can beat thermal paste, you can also hinder cooling performance significantly with them, or just not do any better than thermal paste.
He asked for the best performance so it's liquid metal. But agreed it's downsides are just too much outside of benchmarking and going for records
 
Back
Top Bottom