So what's the best thermal paste in 2017?

grizzly kryonaut

/thread

It's still mayonnaise.

Protip: Make mayonnaise yourself without the vinegar for a more effective thermal paste.

>>59212970
I used this, good price/performance

>>59212970
Literally any paste.

I have Ceramique or something because it was available in a reasonable bulk size for cheap.

>>59212970
Arctic Silver or fuck off.

The shit that comes pre-applied on the stock cooler.

>>59212970
Kryonaut is at the top these days.

Geild GC extreme is also still very good.

>>59212970
I don't think it changes much year over year

http://www.tomshardware.com/reviews/thermal-paste-performance-benchmark,3616.html

they are all the fucking same goop is goop

File: zalman stg1.jpg (21KB, 400x337px) Image search: [Google]

21KB, 400x337px

zalman nail painter

File: indegfdgdfx.jpg (6KB, 275x183px) Image search: [Google]

6KB, 275x183px

>>59212970
prolimatech

its so effective it substitutes an atermarket cooler

>>59214217
>prolimatech
expensive as fuck, while performing worse than shit like gelid and cooler master x1

>>59212970
MX-2

File: 1484910988400.jpg (2MB, 2448x3264px) Image search: [Google]

2MB, 2448x3264px

>>59212970
Just improvise when you cant find your tube of TG-7 grease.

Semen

What would the best thermal paste actually be if money were no object?

If I could buy a paste with platinum and gold in it, would that actually work well?

>>59216010

>>59215994

>>59216029
I could get semen for free though

Is IC Diamond still good?

>>59214255
MX-4

I've had a tube of OP's pic for years now. With my next build I'll use it again.

You only buy these things once in a lifetime as far as I'm concerned.

>>59216031
Yeah you can, sexy. ;^)

>>59216010
Platinum and gold are actually poorer heat conductors than copper and silver.

Theoretically the best you can possibly do is finishing the surface of the chip and the heatsink so microscopically flat that you need no thermal grease. I read about a guy who came close to that once. He was an astronomer and ground telescope mirrors to within a micron of perfect in his spare time. You know you're close when the heatsink and the heatspreader stick to each other when pressed together dry, because air can't get in. He used a tiny drop of thin oil as grease, I think, because the grains of metal in conventional grease were larger than the deviation from perfectly flat in both the spreader and sink.

>>59212970
Definitely not semen.

came_on_my_prox.jpg

>>59216075
Why even bother?
Solder the heatspreader to the heatsink and have them permanently connected.

Or use refrigerator cooling and keep your CPU at a constant room temp + 10 degrees.

without paying more than 15 jewros?
nt-h1

>>59216110
>>Why even bother?
>Solder the heatspreader to the heatsink and have them permanently connected.
Well one, to have them not permanently connected. Two, a pair of perfectly flat surfaces properly mated with no solder will be thermally superior, since you're removing a junction between two materials.

>>59216075
why haven't they ever created a cpu where the heatsink is actually connected to it as one solid piece?

>>59212970
NT-H1
cheap, thermally stable for a _long_ time, not made out of exotic shit.
only downside is that you're not supposed to use it for direct die cooling (no delids)- heard it's got a pump-out effect over time.

>>59216158
Well the heat spreader is soldered to the actual CPU die. Combine the heat spreader and the cooler into a single entity which is then soldered to the CPU die itself.
Instead of a micron perfect match, just have it a solid piece of copper.

Then the problem becomes connection with the die itself- a perfect fit would be tricky.

>>59216161
Because everyone wants different kinds of heatsink designs. OEM desktops want the cheapest thing that can cool the CPU quietly at stock speed. Enthusiasts want giant tower coolers with heatpipes and much bigger capacity. Rack servers want short-height sinks with the fins pointed a particular way. etc.

>>59212997
/thread
Second best is Noctua NT-H1.

>>59213163

THIS GUY IS CORRECT.

File: PICT0010.jpg (311KB, 1600x1064px) Image search: [Google]

311KB, 1600x1064px

Liquid Ultra you plebs.

>>59216258
There's no purpose for this unless you delid and do the same under.

File: f11c1ca2a55927012f6fbb9a4138e9bf.jpg (87KB, 740x494px) Image search: [Google]

87KB, 740x494px

>polishing heatsinks
>not watercooling the die directly

>>59212970
whatever you have is the best

seriously, it doesnt matter, as long its meant to be used as thermal paste it will do the job just fine

>>59213296
Holy shit. According to this, Coolaboratory Liquid shit has an order of magnitude better thermal conductivity than anything else people are mentioning here. Why is no one using this?

>>59216385
Speaking of which, why doesn't Intel make it properly conduct heat without the need for delidding? It's been how many years?

>>59216432
Redpill me on this.

>>59216447
>why doesn't Intel make it properly conduct heat
I've heard one excuse about soldering not being possible because of the die size, but that still doesn't excuse not using liquid thermal paste for example.

>>59216468
stupid bullshit done by overgrown manchildren who don't have a real hobby so they can post pictures online

I was thinking of buying a tube of thermal grizzly kryonaut now that the stuff that came with my 212 evo has run out. Is it worth it? Any better suggestions?

>>59216492
Bluepill?

For the CPU?
None of them.
Use Indigo Xtreme/Maingear EPIC

>>59216468
>>>/pol/

>>59216492
It was a research experiment by a university, calm down.

>>59216447
>Why is no one using this?
It's expensive, it's application is delicate since if you are retarded and smear it on your mobo it will be kill, and it usually glues the heatsink with the IHS making them impossible to separate
This means that you cannot replace your CPU without replacing your heatsink, and also makes it really hard to remove the CPU without destroying the socket

>>59214154
This goyim is COREECT

>>59214154
so after you shave your legs and put on your programming socks you can paint your nails with thermal grease?

>>59216447
It's a bit bothersome to apply and remove.
>>59216563
Pro is the real bitch, Ultra can certainly be removed with some effort. I use lighter fluid and the rubbing thingy supplied in the packaging.
>>59216385
Better thermal paste is better thermal paste, quite simple. It pays itself back by allowing a cheaper cooler (Mugen 4 in my case) perform as well as twice as pricey Noctuas.

kpt-8

>>59213163
Artic silver!! But you have to let set for like 8 days I think before you get maximum performance?

>>59216784
And you have to reapply it regularily.
Not worth the bother.

>>59216784
Why are you faggots so excited about some fucking thermal paste?

kys

>>59216563
>makes it really hard to remove the CPU without destroying the socket
AMD does not have this problem, like at all. I've never heard of pins breaking off because you've lifted the CPU by the heat sink.

>>59213163
>>59216246
>>59216784
Arctic Silver isn't even close to the best any more. Ten years ago, maybe. There are cheaper, better options with no curing time these days.

>>59216563
>Reads up on Galinstan
Weird shit. Apparently the gallium makes it corrode almost every other metal.

PK-3 has served me well

KPT-8, used by soviets. Can withstand a nuclear holocaust.

>>59216193
you know what's biggest problem when welding? air bubbles, it's not that easy to solder to flat surfaces together as it sounds

Dunno but I used Arctic Silver on my GPU recently and dropped temps from 77ºC to 60ºC on load. That shit's awesome.

>>59212970
Pre-applied stock compound because it never dries out.

File: 1486338088490.jpg (369KB, 1000x1000px) Image search: [Google]

369KB, 1000x1000px

>>59212970
>2017
>still using the thermal jew