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

[DSzymborski]

kingston nv2 2tb these are rubbish nvme drives i bought two thinking will use them for storage dont recommend them without any heatsink their temps were 50C while idle. with my own custom garage made $10 heatsink plus $2 graphite grease even these nv2 drives when 100% load never go over 40C.

so beat that

If you actually fabricated your own heatsink, it's not really a $10 heatsink, is it? That is, unless you're valuing your time and effort at $0, which would be an oddly low valuation to give one's own skills.

Of course, the real question is what your temperatures were versus your control runs, when you tested other pastes in identical workloads with identical hardware and ambient temperatures.

 

[WyattStappErp]

Yeah graphite is electrically conductive so I hope this guy used conformal coating to protect the exposed capacitors, etc.

you just don't put it over capacitors all you need a thin layer over cpu gpu or else you need. the advantage of graphite is that its thermal conductivity is around 1800 W/m-K vs paste you used 10 W/m-K 👍

 

[WyattStappErp]

If you actually fabricated your own heatsink, it's not really a $10 heatsink, is it? That is, unless you're valuing your time and effort at $0, which would be an oddly low valuation to give one's own skills.

Of course, the real question is what your temperatures were versus your control runs, when you tested other pastes in identical workloads with identical hardware and ambient temperatures.

how should i value in $ time i spent doing something i like ? its actually fun making something myself.

as you can see on the hwnfo pic there are x two kingston nv2 2tb these are really cheap and nasty when it comes to temps. bought them thinking i dont need anything more expensive for storage/
installed without any heatink and the temps while idle were staying at around 50C then went to store brought something similar to this

temps were still the same around 50C idle went to other store and bought

made of aluminium coated around 7cm tall by 2m long and 1.5mm thick, it looked like i could use it for a heatsink, all i had to do is to cut it into 10cm pieces polish bottom edge of it to remove protective coating also removed red paint between fins of a heatsink. coated bottom of alu profile with graphite paste and inserted two of these between heatsink fins . used few drops of superglue to make them stay together .

and viola temps of these nasty Kingston nv2 drives went down to 28-30 C
and they dont go above 40 even when fully loaded.

once more there are only two fans inside case one for the cpu and one case fan

 

[The Historical Fidelity]

you just don't put it over capacitors all you need a thin layer over cpu gpu or else you need. the advantage of graphite is that its thermal conductivity is around 1800 W/m-K vs paste you used 10 W/m-K 👍

Until thermal expansion and contraction along with gravity slowly migrates the paste out of the heat spreader-cold plate gap and shorts a capacitor millimeters away. Not a risk I would take given most of us have a partner to borrow nail polish from. And by the way, graphite may have that thermal conductivity, but that paste is majority a carrier agent which is a terrible thermal conductor so I’d say more like 18 W/m-K. Still better than ceramic based pastes, but not a factor of 100 better like you claim 👍

 

[WyattStappErp]

Until thermal expansion and contraction along with gravity slowly migrates the paste out of the heat spreader-cold plate gap and shorts a capacitor millimeters away. Not a risk I would take given most of us have a partner to borrow nail polish from. And by the way, graphite may have that thermal conductivity, but that paste is majority a carrier agent which is a terrible thermal conductor so I’d say more like 18 W/m-K. Still better than ceramic based pastes, but not a factor of 100 better like you claim 👍

You do realise that thermal expansion and contraction in graphite doesn't happen unless you heat it to 3000C right? I think it's more like 180 w/M-k plus you can always buy more graphite power and mix it with your graphite paste if you want it to be almost 1800 w/m-k.

Plus no one is forced to do it you can always buy ready made cpu paste from store, I just shared my own experience.

 

[The Historical Fidelity]

You do realise that thermal expansion and contraction in graphite doesn't happen unless you heat it to 3000C right?

You do realize all materials expand and contract at the smallest thermal increments??? It’s basic physics. It may not be noticeable until 3000C as you say, but atomic vibration amplitude is directly affected by even the smallest changes in thermal energy. The higher the vibration amplitude, the more space taken up by each atom, even carbon atoms in a lattice like graphite.

Also graphite paste is a terrible thermal conductor for pastes which is why no reputable company makes a graphite based paste, not only is thermal conductivity drastically reduced by the carrier agent, but paste cannot control the orientation of the graphite lattice. Since the graphite is randomly oriented in its high conductivity (in-plane) and poor conductivity (through-plane) orientations, the effective thermal conductivity is significantly less then you think. There is a reason why reputable companies only use graphite in pads. This is so they can control the lattice orientation to ensure all of it is in an in-plane orientation.

And simply adding more graphite past the optimal ratio is actually worse for effective thermal conductivity due to the paste no longer being able to spread as thin and being unable to remove trapped air from the mixture due to too high viscosity. Air is the worst thing to add to a thermal paste as it basically has no thermal conductivity vs the other ingredients in the paste.

 

[WyattStappErp]

You do realize all materials expand and contract at the smallest thermal increments??? It’s basic physics. It may not be noticeable until 3000C as you say, but atomic vibration amplitude is directly affected by even the smallest changes in thermal energy. The higher the vibration amplitude, the more space taken up by each atom, even carbon atoms in a lattice like graphite.

Lol OK 👍

 

[The Historical Fidelity]

Lol OK 👍

Phd in Biochemistry bub!
But you go ahead and believe you found the jackpot that no one else has, a paste made by a no name company, using an active ingredient that is incredibly picky of its orientation in a medium that does not allow for any control for orientation. Science is not on your side, but who needs science right?👍

 

[WyattStappErp]

Phd in Biochemistry bub!
But you go ahead and believe you found the jackpot that no one else has, a paste made by a no name company, using an active ingredient that is incredibly picky of its orientation in a medium that does not allow for any control for orientation. Science is not on your side, but who needs science right?👍

Lol 👍 i been using graphite grease for a year now haven't replaced any and it seems it does work. It's not no name company because of its graphite and silicone that's all simple to make and extremely cheap. This is best material currently known actually has great heat transfer properties the older it gets it's properties are even better. I won't argue all I did was to share what I learned myself. Go to YouTube and find video showing how to use bike or car graphite grease against pc heatsinks there's none, there are using ketchup butter tooth paste and all that crap so people stick to buying $100 arctic silver 10w /1g instead saving money. I been using lots of different makes and then it came to me i can try something new. I did and it works 👍

and as i said in my first post

beat that with two fans inside your case, one case fan one cpu fan, case fan 450rpm, cpu fan 300rpm, room temp 24C, if you do I will give you $1k.

see Kingston kc3000 in hwinfo ? thats c drive not even once went above 30C normally 25C or same as room temp when idle.

 

[The Historical Fidelity]

Lol 👍 i been using graphite grease for a year now haven't replaced any and it seems it does work. It's not no name company because of its graphite and silicone that's all simple to make and extremely cheap. This is best material currently known actually has great heat transfer properties the older it gets it's properties are even better. I won't argue all I did was to share what I learned myself. Go to YouTube and find video showing how to use bike or car graphite grease against pc heatsinks there's none, there are using ketchup butter tooth paste and all that crap so people stick to buying $100 arctic silver 10w /1g instead saving money. I been using lots of different makes and then it came to me i can try something new. I did and it works 👍

Cool story bro, grifters gotta grift I guess. Good luck with that

 

[WyattStappErp]

Cool story bro, grifters gotta grift I guess. Good luck with that

no need to be rude 👍

 

[The Historical Fidelity]

Lol 👍 i been using graphite grease for a year now haven't replaced any and it seems it does work. It's not no name company because of its graphite and silicone that's all simple to make and extremely cheap. This is best material currently known actually has great heat transfer properties the older it gets it's properties are even better. I won't argue all I did was to share what I learned myself. Go to YouTube and find video showing how to use bike or car graphite grease against pc heatsinks there's none, there are using ketchup butter tooth paste and all that crap so people stick to buying $100 arctic silver 10w /1g instead saving money. I been using lots of different makes and then it came to me i can try something new. I did and it works 👍

and as i said in my first post

beat that with two fans inside your case, one case fan one cpu fan, case fan 450rpm, cpu fan 300rpm, room temp 24C, if you do I will give you $1k.

see Kingston kc3000 in hwinfo ? thats c drive not even once went above 30C normally 25C or same as room temp when idle.

That’s normal bro, let’s see the temp when heat is actually being produced

 

[WyattStappErp]

That’s normal bro, let’s see the temp when heat is actually being produced

What normal?

 

[WyattStappErp]

Whats normal "bro" ?

 

[WyattStappErp]

here are temps while copying 100gb file from 1 nv2 to other nv2

here are temps right after crystaldisk 64gb test against one nv2

 

[DSzymborski]

7% CPU and 0.8% CPU are hardly the most taxing benchmarks.

 

[ReveurGAM]

Yeah graphite is electrically conductive so I hope this guy used conformal coating to protect the exposed capacitors, etc.

Do you mean like petroleum jelly?

 

[The Historical Fidelity]

Do you mean like petroleum jelly?

If you like a mess lol, no I mean kapton tape, nail polish, a substance known as conformal coating, etc.

 

[WyattStappErp]

7% CPU and 0.8% CPU are hardly the most taxing benchmarks.

this was ssd test 👍

 

[Black6spdZ]

why are actual CPUs still being used on these TIM tests?! They have varying loading that can hardly be repeatable and pretty much become obsolete or not even applicable with a newer or different CPU. INSTEAD, use a standard heat loads "eg 50w, 100w, 200w, 250w" with a resistive heating element with standard contact sizes 1000 mm2, 1250mm2 etc and measure delta T < this is the only way to get results that are comparable over time

 

[ReveurGAM]

why are actual CPUs still being used on these TIM tests?! They have varying loading that can hardly be repeatable and pretty much become obsolete or not even applicable with a newer or different CPU. INSTEAD, use a standard heat loads "eg 50w, 100w, 200w, 250w" with a resistive heating element with standard contact sizes 1000 mm2, 1250mm2 etc and measure delta T < this is the only way to get results that are comparable over time

Well, I have two reasons I don't:
1) I can't afford to buy such a device from LonGwin, etc.
2) I have no idea how to design and build such a device, although I'd love to have one.

 

[Ser Menalak]

This article is disingenuous and needs to be taken down or updated. You CANNOT say these are the best thermal pastes for 2024 when these tests were clearly done years ago and you don't include any of the more recent and better performing pastes.

 

[USAFRet]

This article is disingenuous and needs to be taken down or updated. You CANNOT say these are the best thermal pastes for 2024 when these tests were clearly done years ago and you don't include any of the more recent and better performing pastes.

"last updated November 22, 2023"

What new ones are missing that you think should be in there?

 

[Ser Menalak]

"last updated November 22, 2023"

What new ones are missing that you think should be in there?

The article says 2024 when you see it from google. The forum post doesn't for whatever reason. Where is Thermal Grizzly Kryonaut Extreme which performs about 2-3c better than standard kryonaut? What about the new Arctic MX-6?

 

[COLGeek]

The article says 2024 when you see it from google. The forum post doesn't for whatever reason. Where is Thermal Grizzly Kryonaut Extreme which performs about 2-3c better than standard kryonaut? What about the new Arctic MX-6?

Title corrected. Content remains.

There will ALWAYS be more products than any review can assess. Maybe these can be considered for a future review.