I personally go back and forth from Thermal Grizzly Kryonaut , Conductonaut, and MX-4 all depending on what I have at the time. AS5 is sort of my last resort because I normally always still have a tube of it laying around.

I will say my first attempt with Liquid metal didn't go so well as a retard I tried to apply while the CPU was still in the socket, It dripped as I came up from putting it on the CPU and went into one of the first memory slot then of course I had a missing stick of memory being detected but luckily it was easy enough to remove out of the memory slot and all was good.

Great review, thanks. I'm surprised how well Gelid Xtreme holds up still. I tried it years ago after delidding my 6700K because it suffered terrible pump out between the core and lid within a day. Gelid Xtreme being very thick took weeks for the same effect to happen and gave better temps too. I think I was using MX-4 before that.

I've only ever used Arctic Silver so seeing where it placed on the chart was.... interesting.

Silva:

@TigTex you bring an interesting topic, I didn't know thermal compounds could "run away" or migrate to the corners worsening performance. Thankfully I've never experienced this. And ya, laptop designs are generally trash, I don't know if it's incompetence or they just fracking with us. I've been using a single 4g MX-4 syringe for the past, I don't know, 10 years? Never had a single issue and I recommend it to everyone who asks. It's a great cost/performance option and I never got the "glued CPU to cooler" situation ever again. PS: never use pre applied thermal compounds: they suck! (I'd say literally, but they are more like glue)

[youtube=MyvDZUGlEQI] Pump out is only an issue if you use the paste on the die. Not when used over and IHS.

SplashDown:

What are you using for paste?

Kryonaut both for air as water on GPU and CPU, takes a bit more work applying it and a day or two to fully settle in.

TigTex:

One extremely important factor with thermal compounds is the pump out effect. Laptops with crap cooling that cause high temperature deltas are extremely prone to that and I'm still having issues finding a thermal compound that doesn't escape from under the heatsink after a couple of months (somes are just days!). For example, I have a Lenovo T490 that only spins the fan after 80ºC (nice design, right? no). The constant heat cicles cause the thermal compound to escape from the CPU die. So I start by having max temperatures of 85ºC and almost no fan spin and 1 week later the temperatures are above 90ºC and the fan spins everytime I do any activity like opening a simple browser tab. When I see the heatsink, the paste has migrated to the corners of the cpu and the die is almost without paste. MX4 suffers a lot from this. I used it on thousands of laptops when I repair them at my shop and I see that almost all of them fail after short term. Coolermaster mastergel pro v2 provided me the best results but it is still not working with my lenovo laptop. Kryonaut was one of the worse, causing thermal throttling in less than a month. Probably one of the best "paste" is the carbonaut because it doesn't suffer from the pump out effect. This issue only affects laptops and gpu's with poor mounting pressure and I've never seen anybody running those tests. Should be easy to replicate (but time consuming): low mounting pressure with high temperature deltas

I haven't really had the need to replace thermal compound on any laptop I've owned for years so I don't have a good recommendation, but have you considered creating a silicone gasket to help keep the compound from leaving? You'll want to make sure it's properly vented of course, but that vent shouldn't be enough to allow for thermal compound to leak.

@Hilbert Hagedoorn , you've got your Hydronaut & Carbonaut pages & photos mixed up: https://www.guru3d.com/articles_pages/guru3d_thermal_paste_roundup_round2_2021,4.html https://www.guru3d.com/articles-pages/guru3d-thermal-paste-roundup-round2-2021,5.html Still reading the review, will edit this post if I've got anything interesting to say about the products or results. EDIT: Thermal Grizzly Kryonaut still right near the top in performance, I'm using it on my CPU and GPU, I've had it on my CPU since 2017 and it's still performing just as well....I think part of the longevity is down to the fact that my liquid metal delidded 6700K doesn't ever go above 65degC, and I remember reading in the literature of Grizzly Kryonaut that the paste won't dry out if kept below 70degC (or was it 75degC)....also the mounting system of the Noctua NHD14 cooler I use is high pressure/flat and stable/secure so that might contribute to the longevity of the Kryonaut paste which is now essentially 4 yrs old without any need of repasting.

I would suggest testing Thermal paste in laptops. Due to the pump out effect (at least that is what der8auer called it), it is probably the most important place to know what thermal paste to use. I personally used 3 different pastes (Arcitc silver 5, thermal grizzly ... I think it was cryo and a standard thermaltake paste) and all of them had the same problem: after a few thermal cycles (like 2 weeks to 6 weeks, 2 in summer, 6 in winter), the temps were very high again. When I went to replace the thermal paste I saw almost none of it was still left on the contact area, it was mostly on the sides of the contact plate/CPU or GPU die. Right now I dont have a laptop but once I do I will have to find a fix for this issue, because my temps right after replacing the TIM were in the high 60s for the GPU and high 70s for the CPU (this was an old laptop, i5-2670m and 570M) and after a few weeks: high 80s ion the GPU and mid 90s on the CPU. And the more wide spread the temps get in a laptop, the higher the chance that something will die quicker, especially in bigger GPU dies. If your Laptop is at 20C when turned off, and at 80+ when on load, that is a rather large variance for the metal parts that will contract and expand due to thermal expansion.

JiveTurkey:

Interesting. Thanks for the tests. I am gladly willing to sacrifice 5c for something that lasts longer. Some of the new pricey hot shots can dry out much earlier. Even heard of some of them needing to be reapplied within months! Using the cryonaught stuff now, so we'll see. But I'll say this for AS5, never had an issue with anything unexpected after pulling a hsf off after 5+ years of use. So far have had an issue with dry paste on mx and the NT. Also recall a review where a paste had performed way below expectations and the review contacted company and got sent a fresh the tube and it made a big difference. I seriously wonder what these benchmarks would like after months plus years and thousands of power cycles. But the effort involved in that dang.

Thermal paste looking "dried out" doesn't mean it stopped working. If thermals do no change, who cares if paste looks dry next time you open system? Lots of thermal compounds can be dry and still work fine.

@MerolaC this is a real issue you'd think someone thought about solving long time ago. I'd say you're better using a different metal (one that expands similarly to silicon) and the usual copper heat pipes on top to carry the heat.

Silva:

@MerolaC this is a real issue you'd think someone thought about solving long time ago. I'd say you're better using a different metal (one that expands similarly to silicon) and the usual copper heat pipes on top to carry the heat.

I don't think the issue could be solved from the cooling portion since it will most probably increase the price of it. The thing is that most manufacturers, use what people call "crappy TIM". The dry looking paste is in fact carbon based with no silicone, so it doesn't pump out and looks like that since day one. Yes, it might not be the most effective paste in terms of thermal transfer, but it will last forever in your CPU die without pumping out.

Would love to see GD900 here on the list. its giving me amazing performance for the money

For years rule of thumb Kryonaut for CPUs and Conductonaut for GPUs. Job done

I prefer just using mx2 - easy to work with, doesn't run, and doesn't go hard for many years.

JiveTurkey:

Interesting. Thanks for the tests. I am gladly willing to sacrifice 5c for something that lasts longer. Some of the new pricey hot shots can dry out much earlier. Even heard of some of them needing to be reapplied within months! Using the cryonaught stuff now, so we'll see. But I'll say this for AS5, never had an issue with anything unexpected after pulling a hsf off after 5+ years of use. So far have had an issue with dry paste on mx and the NT. Also recall a review where a paste had performed way below expectations and the review contacted company and got sent a fresh the tube and it made a big difference. I seriously wonder what these benchmarks would like after months plus years and thousands of power cycles. But the effort involved in that dang.

nth performs very badly over time dc1 was the only one I haven't seen any deterioration in 3 years time.

TigTex:

One extremely important factor with thermal compounds is the pump out effect. Laptops with crap cooling that cause high temperature deltas are extremely prone to that and I'm still having issues finding a thermal compound that doesn't escape from under the heatsink after a couple of months (somes are just days!). For example, I have a Lenovo T490 that only spins the fan after 80ºC (nice design, right? no). The constant heat cicles cause the thermal compound to escape from the CPU die. So I start by having max temperatures of 85ºC and almost no fan spin and 1 week later the temperatures are above 90ºC and the fan spins everytime I do any activity like opening a simple browser tab. When I see the heatsink, the paste has migrated to the corners of the cpu and the die is almost without paste. MX4 suffers a lot from this. I used it on thousands of laptops when I repair them at my shop and I see that almost all of them fail after short term. Coolermaster mastergel pro v2 provided me the best results but it is still not working with my lenovo laptop. Kryonaut was one of the worse, causing thermal throttling in less than a month. Probably one of the best "paste" is the carbonaut because it doesn't suffer from the pump out effect. This issue only affects laptops and gpu's with poor mounting pressure and I've never seen anybody running those tests. Should be easy to replicate (but time consuming): low mounting pressure with high temperature deltas

Plenty of people on the laptop forums have done tests like this. You just don't see them around here since everyone else cares about desktops and video cards. But I think this is what you were looking for. ( http://forum.notebookreview.com/threads/which-thermal-paste-to-buy-and-apply-traditional-and-liquid-metal.806840/page-48#post-11084441 ) (if that link gets truncated: (forum.notebookreview.com/threads/which-thermal-paste-to-buy-and-apply-traditional-and-liquid-metal.806840/page-48#post-11084441") Try Thermalright TFX? I think this is one of the best pastes for something with weak contact, with IC Diamond probably being the best. But keep in mind literally everything you would use would degrade if the contact is that bad (except IC Diamond), and graphite pads like these wouldn't help you. https://www.amazon.com/IC-Graphite-Compression-Thermal-40X40mm/dp/B08XNYHYCH/ (I believe this is Panasonic EYG-R, which is newer than the older EYG-S). Compare this to https://www.amazon.com/Innovation-Cooling-Graphite-Thermal-Pad/dp/B07CKVW18G (this is EYG-S). I bought both of the panasonic brands directly from digikey, and I still have the original IC packs from when I was testing for them. Also see if you can do a contact pressure test on the laptop with contact paper https://www.innovationcooling.com/product/ic-contact-test-analysis-kit/ (this is the same material as Fujifilm "Ultra Low" Prescale (25-85 PSI), which you can request a sample of (politely) from sensorprod.com). There is also Extreme low prescale, which is I think 6-25 PSI. If you determine the CPU or heatsink is convex, you can try to do a mild sanding with 600 grit-3000 grit sandpaper (sand only enough to flatten the convexity slightly). Then you can try to reapply.

Dragam1337:

I prefer just using mx2 - easy to work with, doesn't run, and doesn't go hard for many years.

For all these years after using God knows how many different thermal compounds MX-2 is still my favorite one. I've spend last tube of MX-2 last year, can't find it here localy past few months so now i'm "forced" to use MX-4 and NT-H1. Shipping MX-2 (same as anything that costs more than $2 to $3) from eBay/Aliexpress is insanely expencive past year or so, so atm I'm waiting for a good local deal for MX-2 than I'm gonny buy supply of MX-2 that'll last me at least 10 more years. But will try MX-5 after this review...

I always use: Thermal Grizzly's Kyronaut. I currently using on 4 of my Titan Xp's: Kyronaut Extreme. I have seen difference from 2c to 7c over the stock paste that it came with, that being said, this was at a glance, as I have not checked what is after hrs after use(Curing), even though Thermal Grizzly claims no curing time, I see otherwise and other sites too: Check this forum, look where the see the testing results after some use: https://www.techpowerup.com/forums/threads/kryonaut-12-w-mk-vs-conductonaut-liquid-metal-73-w-mk.229167/ AS5 is good for systems that are not going extreme with the overclocking, such as: Voltage, Speeds, etc. One thing I love about using AS5, spreads evenly, no separation, but it is conductive, Never had a problem with AS5, not to my knowledge. I have personally tested AS5 on my hottest gpu while I was under 4-Way Sli(Believe while using Spacers), and this was(If I remembered correctly, last summer), it was hovering to 90c, it wanted to go higher, so I backed out of it, Next, after cooling down the system for a bit, changed the Thermal paste to my personal Favorite: Thermal Grizzly's Kyronaut, I have seen a 12c drop easy, Absolute staggering difference, and these were at a glance tests, these test were not conducted in a long hrs worth of test. Now, I haven't personally tested the difference between Kyronaut and the Extreme version, I would like to see differences on my own end, I think within long hrs of use, the difference would probably be almost nil, I wouldn't expect much difference, maybe a 1c or two tops, Haven't had an issue with it yet. Thanks for sharing this.

@A M D BugBear so your surprised a 21st century Paste outperforms a +6y "old" paste (and according to Arctic not recommended/replaced by MX4/5) ? ok 🙄. i dont mind paying a bit more, and did see temps drop a few *C on LM, but i wont pay multiple times more, for only a few degrees difference. MX is 7$ for 4g, and anything from TG is at least 50% more (while up to 4 times less amount), without being "twice as good". Another reason for me to stay with MX5, was how easy it spread, while not getting "repelled" from oily "residue" (of previous TP applied).