Thermal pastes don't change their performance based on whether they're used on a CPU, GPU, NVMe or chipset. The rate at which they transfer heat remains the same so the best thermal paste for a GPU would also be the best thermal paste for a CPU. Let's be logical here, thermal paste doesn't care where the heat is coming from, it just does its job.
Now, in all truth, there's very little difference between thermal pastes. I coated the exposed GPU die of an RX 580 2048SP with HY-510 thermal paste and it worked just fine. I used the same thermal paste to coat the exposed die of my old Acer Craptop's A8-3500M APU. Again, no problem (and I did it over a year ago).
Just look at this chart from this Tom's Hardware's article:
The difference across ALL of the
actual thermal pastes from what could be called the best to what could be called the worst is 10.3°C. As you can see, Arctic MX-6 isn't anything special, in fact, it's ranked 28th out of 51 which makes it technically below average (the average would be ranked 26th). Having said that, there's nothing wrong with Arctic MX-6, it's good thermal paste and will do the job just fine so if you were to choose it, you wouldn't have any problems.
People often over-emphasise the importance of thermal paste because the truth is that the chip being cooled and the cooler being used for it are both
FAR more important when it comes to temperatures than the thermal paste used.
Keep in mind that article uses a worst-case scenario running an all-core workload on an Intel i9-14900K, the only consumer-grade CPU ever made that out-draws and out-heats the old AMD FX-9590. My CPU is an AMD Ryzen 7 5800X3D and I'm just running a standard AMD Wraith Prism air cooler (which cost me nothing). According to this chart, HY-510 is technically one of the worst out there with a temperature of 67.5°C but that's still 32.5°C
below the i9-14900K's TJ Max threshold of 100°C.
Someone once asked me why I bought such "low-performance" thermal paste and my answer was the same as the answer I'm giving you. It doesn't really matter what kind of thermal paste you use (unless you're some insane overclocker) and someone on eBay was selling
TEN 30g big tubes of it for $25CAD. That's right, I bought 300g of thermal paste for 25 Canadian Dollars. There's no question that I'll
never have to buy thermal paste again. A couple of friends even bought tubes off of me ($5 each) so I really bought eight 30g tubes for $15CAD which is an even better deal and I probably still won't even be able to use a third of it. To give you an idea of how amazing a deal that was, $25CAD would only be enough to buy
a 4g tube of MX-6 from newegg.ca. Yeah, I'll take the 300g of HY-510 over 4g of MX-6 every day of the week and thrice on Sunday.
I would advise against using liquid metal pastes though because even though they do out-perform the more common (and less expensive) thermal pastes, they can be hazardous to your circuits. This is because liquid metal pastes are conductive and can cause a short between parts on a PC board (motherboard, video card, expansion card, etc.) and fry the circuits if you use too much and it spills over. The slight temperature difference isn't worth the risk as far as I'm concerned and even though I would be qualified to use them correctly, I still wouldn't take that chance.
In any case, we can see that even the control substances (ranch sauce & toothpaste) don't approach the TJ Max of the CPU. This means that as long as you get a purpose-made thermal, even the cheapest stuff you can find, you won't have any problems.
Everything will be just fine, I guarantee it.
Facebook
Twitter
Instagram