Thermal paste and Computer Building

[Thunderball]

I have never seen or heard of anyone doing this.

You probably never heard about anyone doing serious overclocking too.

[the game]

This is what I use.

http://www.microcenter.com/product/4...ermal_Compound

[SoulForge]

You probably never heard about anyone doing serious overclocking too.

My 4770k is OC'd to 4.5. And has been since shortly after its launch when I purchased it. I check the temps every few months and it hasn't changed.

Three years later. No paste change.

[Thunderball]

My 4770k is OC'd to 4.5. And has been since shortly after its launch when I purchased it. I check the temps every few months and it hasn't changed.

Three years later. No paste change.

Told you why, you probably applied a lot of paste so it's still good. You could probably get better temperatures (if your heatsink is heavy enough) but it's not important for you because 4.5Ghz is not serious OC.

[SoulForge]

Told you why, you probably applied a lot of paste so it's still good. You could probably get better temperatures (if your heatsink is heavy enough) but it's not important for you because 4.5Ghz is not serious OC.

3.5 to 4.5 is not serious? Not going for world records here.

[Deleted]

Never had to worry about TP quality, just take whatever. It's always nice to have a bit lying around, never know when a relative with 6 years old laptop might come by.

[Thunderball]

3.5 to 4.5 is not serious? Not going for world records here.

Usually not enough to push it to the temperature limit where every degree matters.

[Lathais]

This isn't true. A good paste can make substantial difference, like > 5 degrees of Celsius or so.

Unless you go with the liquid metal stuff, not really.

Not even a 3 degree difference, unless you want to count the much much more expensive liquid metal options. Even at that though, they are only offereing about another 2-3 degrees over the best pastes and only 11.5 degree difference from denture cream. Personally, I did use the Coolaboratory liquid pro on my current builds, just for the hell of it, but at the temps my CPU stays at, even under stress test, I'd be fine with denture cream. Toothpaste might be getting a little hot under stress test, but appears as though it would be fine under normal usage. So no, not really a substantial difference.

- - - Updated - - -

3.5 to 4.5 is not serious? Not going for world records here.

Considering the default turbo boost is 3.9, not 3.5, it's only about half the OC you think it is, and yeah, not that much. 4.5 is a pretty average OC for that CPU, not serious at all.

[ifrah]

Poster you quoting is corrent, it's recommended to reapply thermal paste on the CPU every year if it has been applied properly (thin layer) and the paste itself applies well. You can apply more paste so that you wont have to change it that often, but your temperatures will be higher.

--- snipped off, misread the post, sorry ----

And the aging. Yes, paste will flow off from there to some extent over time, while a good part of the solids will not go anywhere. Thermal cycling accelerates this to some degree. Now the question is, when does this effect become relevant. And the answer is: when the system starts to overheat / become unstable / throttle. If none of that happens, few degrees here and there make no difference. And since judging by OP's CPU, which is not a "K-model," any serious overclocking is not part of the plan. In the non-overclocked systems I've had or heard of since the 90s, be the system 1 or 5+ years old, they have had no overheat issues whatsoever if they have had none to begin with.

What comes to serious overclocking where every degree counts, I'd still however recommend the same: If the system shows no signs of overheating, why bother?


edit: misread the quoted post. removed irrelevant part

[Lathais]

Applying more paste makes it better? How about no? The paste is no magic stuff that makes everything so much better. In the best ideal scenario you would not want to ruin the thermal conductivity with anything. However, as we know from high school, every surface is uneven. And ideally you want to fill those cavities that will form between these surfaces, not to create some wicked sandwich.

And the aging. Yes, paste will flow off from there to some extent over time, while a good part of the solids will not go anywhere. Thermal cycling accelerates this to some degree. Now the question is, when does this effect become relevant. And the answer is: when the system starts to overheat / become unstable / throttle. If none of that happens, few degrees here and there make no difference. And since judging by OP's CPU, which is not a "K-model," any serious overclocking is not part of the plan. And in the non-overclocked systems I've had since the 90s, even HW running 7+ years have had no overheat issues whatsoever if they have had none to begin with.

I do not see anywhere where he said the word better. He said that it lasts longer but gives higher temps.

[Cilraaz]

Poster you quoting is corrent, it's recommended to reapply thermal paste on the CPU every year if it has been applied properly (thin layer) and the paste itself applies well. You can apply more paste so that you wont have to change it that often, but your temperatures will be higher.

i5 2500k @ 4.6GHz using the same Arctic Silver 5 since I bought it in 2011.

Every year is excessive. Every few years? Perhaps.

Also, this is in a thread where the OP has a non-K CPU, so there's no 5GHz+ overclocking going on.