Arctic Silver 5 vs. IC Diamond

It's really setup dependent. I've had AS5 and IC7 loose their ability to cool after a few ventures into 90 degrees during benchmarking. After that they lost their ability to cool even normal temps.
For non performance related tasks and gaming AS5 and IC7 are perfectly fine and would last long enough - more so IC7 than AS5. But for performance related tasks, overclocking, CPU intensive gaming I'd go CLU everytime. It's magic on CPU's with no IHS cops everything you can throw at it on the chin.

 

No... Noctua NT-H1 is not a good paste is my opinion. What processor do you use in your GT70 barebone? Is your processor overclocked like mine? Have you tested the processor to the limit in diff benchmark tests? If you run your processor and gpu overclocked in benchmark tests you're going to see the difference between Noctua NT-H1 and a better thermal grease. Guaranteed.
Read also this;
http://forum.notebookreview.com/thr...ips-before-i-start.741745/page-5#post-9520969

 

I see. I'm not an overclocker person (not for the CPU anyway), so on that circumstances it can lose its one day efficiency during extensive loads with overclocked. Just like most of thermal pastes out there. CLU performs excellent, except I didn't trust myself not to spread it accidentally to electronic components because it would end up with terrible results as CLU is indeed conductive. Most of people don't choose CLU because of this. It would have been perfect if there is a non-conductive version.

It is i7 4710MQ. No, it is not overclocked like yours, the maximum GHz I went into is 3.7 for one core and not for a long time. Most of the time my CPU is locked it to 3.0 GHz as it is sufficient for me. But I've tested with its turbo clocks (3.5 GHz) a week ago in 3DMark test and Prime95. The results were about the same as in March and there was not a big difference when you remove the ambient temp effect.

I saw Mr. Fox's review. I saw another review states that Noctua NT-H1 can indeed lose its one day efficiency after months of use when you applied it to bare IHS. It was the best performing thermal grease in the test among many on bare IHS before losing its one day efficiency, as I remember even beating CLU and IC Diamond (can't find that thread now, I think it was on overclockers.net). I would do repasting each six-seven months and it would be enough in my situation. If you are experienced and willing to take the risk, then I would recommend CLU as it seems does a better job in a long term performance and durability.

 

Using already CLU. No(zero) change in temperature after over 1.5 years use on my Mx processor... There is no better paste for laptop processors which normally has a Inadequate cooling. Look at this. Check out the thumbnails in the link. PK-3 one of the better thermal grease, but falls short against Cool laboratory Liquid...
http://www.guru3d.com/news-story/intel-skylake-de-lidded.html

 

Interesting thread.

Many folks have no idea what "TIM" is, nor would they ever change or re-do it.

Is there a study that shows which one, once properly applied, would hold up best under warm temps and last the longest?

 

There are plenty of tech websites out there that have compared thermal pastes. The general summary of all of those studies are that name-brand thermal pastes are all better than generic cheap thermal paste by a significant margin (5C - 10C). But they are all about equivalent to each other, within a margin of error (0C - 2C).

So it really doesn't matter what thermal paste you use (as long as you have some kind of high-end name-brand stuff) since you won't notice any real-world difference in any of them. The real difference at that level comes down to cost, ease of application, ease of clean-up, and electrical conductivity.

 

Thanks kent - I understand that.

What I'm looking for would be studies that show how long a proper application lasts.

In other words, if three different products give essentially the same initial results but after, say, 18 months two of those products must be replaced, I'd like to know about that third one.

 

If the thermal paste doesn't specifically state a shelf life or expiration date, then you don't need to worry about it. Most of the stuff in the pastes are chemically inert.

Some people replace their thermal pastes every 2 years. But they do that out of an OCD behavior, and not to actually solve or prevent a problem. If the paste is properly applied and doing its job, it should last indefinitely.

Sent from my XT1575 using Tapatalk

 

Actually, I would argue that habitually replacing thermal paste is a BAD thing.

There's a decent chance that the thermal paste is stuck on so well to the CPU / heatsink, that the act of removing the hestsink actually poses a risk of causing damage to your hardware (rip off pins, rip off socket, rip off heat spreader, etc).

Never in my entire life have I ever replaced properly-applied thermal paste because it "dried out" or "went bad". Every time I have re-applied was to either correct a mis-application, or to replace a heatsink.

Sent from my XT1575 using Tapatalk

 

@TBoneSan @D2 Ultima @Mr. Fox @Matrix Leader @TomJGX
What you say is erroneous/not correct... Many of those studies you talking about are mostly tests done with desktop processors. It is not possible to compare these tests with tests done on many laptops who have different cooling(some have good and some very bad cooling). I can guarantee you that there is a bigger difference between high-end name-brand greases. Don't forget that the cooling in many laptops is much worse than in desktops. Heat dries out the thermal paste a lot faster than you think... Especially if the cooling is really bad. The main difference lies in the lifespan of paste. This is actually very important. It does not help with good results from some of these so-called high-end name-brand stuff if it turns into a miserable results after one week or 2 month. Have you by the way seen the comparison between IC7 and the highly acclaimed Noctua Nt-H1 in the link from @Mr. Fox ? http://forum.notebookreview.com/thr...ips-before-i-start.741745/page-5#post-9520969 Advise you to test different high-end name-brand greases on real high end laptops with a decent/high overclock of the processor and GPU. Think you will quickly change your mind . I have been there. Have tested. Don't forget that I have a extreme laptop processor that smooth running with over 100w in bench tests. This is twice the max stock TDP on laptop processors. Guaranteed a difference.

 

Fair enough to point out.

Thermal pastes being used in laptop environments (which often skirt the very limits of how much heat the cooling system can dissipate) are probably much more important, and sensitive to quality and performance, than in desktop environments.

 

Master Papusan, a word of advice please!

As I use a Sandy-E Desktop Proc with no Chance of delidding I am stuck (best?) with ICD or GC Extreme, right? Currently I am using ICD and I would be satisfied with temps, if I wouldn't get different temps on the cores.

Idle: (22° ambient, stock settings)
Core 0 - ~40°
Core 1 - ~33°
Core 2 - ~33°
Core 3 - ~42°
Core 4 - ~43°
Core 5 - ~35°

As far as I can say the TIM spread evenly all over the Proc. So I presume I either did a bad pasting Job anyway or the heat sink might be uneven. Is there also an indicator, if the heat sink is applying proper pressure on the proc? By let's say where the TIM is located mainly?

- Most of the TIM on heat sink?
- Most of the TIM on Proc?
- Equally on both -> heat sink and Proc?

I can provide some pics at the Weekend if needed. Would be great, if you could throw your 2 Cents in!

 

@CaerCadarn It's normal with temp difference between coldest and warmest core ofthen about 6 degrees Celsius. Your processor is slightly in the upper edge with 10 degrees. What is the difference between coldest/hottest cores in stress test? Same difference as idle? If you're not satisfied, I recommend you do it all over again. Heat the thermal grease in a bag and put it in a cup of 55 degree water a while before application(only if you have not done this already). Especially important with ICD or a any other thick paste's. Use minimal paste if you have a good level heatsink and proper pressure on the proc. If the heatsink is level and the pressure of the heatsink is good then you should got a good result anyway. Is all this in place then I recommend to test with Liquid ultra. You do not need delidding the processor to get a very good result with Liquid ultra(especially under maximum load). High pressure and a level surface is perfect match for use of CLU. Run and buy now... Think some people say that it is possible with delidding Sandy-E Desktop Proc . But I don't recommend it. Remember Sandy-E Desktop Proc use fluxless solder who generally has a much improved thermal conductivity, Typically in the range of 80W/mK, whilst a standard TIM paste has only 5W/mK.
Links http://hexus.net/tech/news/cpu/39369-intel-cuts-corners-ivy-bridge-thermal-interface-material-tim/
http://techreport.com/news/25004/report-delidded-ivy-e-processor-has-solder-under-its-heatspreader
Some run5ghz. See the temp...

 

Tack sa mycket för kommentera! Will try that out soon.

Just one more question: Is CLU also recommended for pasting GPU's? Wasn't sure about that, cause I found contradictory information to this topic.

 

CLU is not recommended for GPU's... It's extremely dangerous... @Mr. Fox has done it before but he's like a god of repasting

Use Gelid GC Extreme for GPU or IC Diamond..

 

Did a quick stress test via CPU-Z on stock levels. Not much headroom left. So I suppose the pasting with ICD was not done properly....

Spoiler: CPU-Z

 

Mr. Fox run i7-4930k overclocked at 4.3ghz on all 6 cores 24/7 in his Clevo P570 "Panther" without heat problems. He has used both ICD and Liquid ultra. Your problem is most likely a bad processor heatsink (to large gap between the die and heatsink). But ask @Mr. Fox about this issue and what temperature you would expect with a proper application of paste on your hardware. Liquid Ultra on the graphics card is quite safe if you do the job properly. Many cover around the graphics die with tape or clear varnish. See the link. http://forum.notebookreview.com/thr...h-alienware-17-r2.782619/page-2#post-10112451

 

@Mr. Fox: One word of advice please. You mentioned using some sort of sensitive pressure film to test the contact between CPU und heatsink on your Panther. Where did you get it? It seems rather unpurchasable....

It's frustrating getting crappy results with ICD on stock. If the heatsink is warped somehow, then I would like to know before applying CLU on it. My intention is to know how good my Sandy-E is overclockable and depending on thus eventually replacing with the 4930k.

 

There is no point in applying Clu if the cpu heatsink is not in top condition. ICD paste is the only and best option if the heatsink is warped. Can you run a Wprime 1024 stress test (select 12 treads)? http://www.putme.net/wPrime155.zip

 

Sorry for the late reply! Here we go:

Spoiler: WPrime1024

Edit: Did another Stress test and what I see makes me start worrying....

Spoiler: Prime95

I am not able to interpret it correctly, other than there is mentioned the "Fatal errors". But anyway this isn't that good, right?

 

Recomend rather to use Prime 27.9 instead of 28.4 on laptops. I am interested in the temperature during Wprime stress test. Wprime is more comparable to real use. Perhaps Mr. Fox can tell you what the processor maximum load temperature should be in your Clevo P570. Easier to find out if the heatsink is good or not when you have something to compare with.
These peak temp results in different stress tests may vary with the cooling and setup you have. Probable tested with a desktop PC. But an ok pointing stick.

 

Remember too that the heatsink pad over the chip for proper CLU installation must be copper. It is highly corrosive to aluminum. As it is now with my NP700G7C I run CLU on both the CPU and GPU. I seem to get about the same results as with ICD that was there prior to the application. I had to try it out though.

 

Thank you for that! Will get back soon with some more comparisons. But what I've seen so far is that my Proc is running about 10° hotter in XTU stress test than that provided in your screenshot.

CLU should be more effective in heavy workloads. But at least, if the possibility of a warped heatsink can be excluded, I'll give CLU a try.

Still not convinced that my paste Job was that bad. Will post a pic soon....

Btw, is this an appropriate tool for testing the heatsink, if it is warped or not?

 

The test temperatures for different bench tests are probable done on a desktop PC. Intended as a guideline on the temp differences between stress tests.
Think this is the right paper for measuring pressures. http://www.sensorprod.com/pressurex-micro-green.php?app=Heat Sink

 

Okay, got the WPrimeTest with temperature now. Dunno why it took that long this time?

Spoiler: WPrime

 

Select an equal number threads as your processor supports in the advanced settings aka 12 threads .

 

Oh man! I shouldn't do stuff when I'm tired!

Here we go again:

Spoiler: WPrime

 

So unfortunately it took a long time to get this pressure film but here we go:

Spoiler: Pressure Test1

That's the result after screwing a little bit "tighter":

Spoiler: Pressure Test2

As far as I can identify it, it seems that the screws apply different pressure as the right ones had lesser "cycles". Here some pics of my paste Job with ICD, which wasn't so bad imo:

Spoiler: Paste job

 

This may be the Explanation why three cores ran hotter as the other ones on idle and under load:

Spoiler: XTU_stock

I will try to find out, if IHS is uneven or the heat sink may be warped. So no CLU yet!

 

It looks like you have the same problem as @Mr. Fox The old heatsink on his P570 had a similar problem as you(warped heatsink). If I remember correct; The issue was solved with a new and better heatsink. He could also use Clu after this replacement.

 

My laptop was reaching maximum of 70°C while few tabs opened and listening to music, I have repasted(after 2 yrs) with ICD and the maximum temperature gone down to 58°. I'm really pleased with the outcome

 

I promise you, that temperature drop is due to the proper (re)application of the thermal paste, and not the brand of thermal paste used.

Sent from my XT1575 using Tapatalk

 

I agree with you, I have a tube of AS5 as well but it's very old(8+ years) so I chose to use ICD as I have no preference.

 

It's a bit of both tbh... Good thermal paste makes the difference... Especially in laptops where quite a few brands pump out easily... No such issues when using the cheap but high quality Gelid GC Extreme and IC Diamond pastes...

 

I would like to see a laptop with hardware temp below 50C on max load
BTW. Welcome to the forum

 

Show me an example too

(Not happening!)

 

i prefer neither of these.

I have actic mx4 and gellid ultra.

stuff that is easy to remove and its easy to apply.

 

GC Extreme is easy to apply and remove and does quite a good job too, much better then the 2 you use ...

Sent from my LG-H811 using Tapatalk

 

I don't think there are many other paste which is easier to apply Also see this below
http://forum.notebookreview.com/thr...x-980-overheating.789202/page-5#post-10218939

http://forum.notebookreview.com/thr...x-980-overheating.789202/page-5#post-10218974

 

In fact, if you warm up ICD a bit, then it is fairly easy to apply too!

 

I know, but I prefer Liquid Ultra . Whatever paste I used earlier, I warmed it up first

 

Yeah, sure! In lovely Norge it's necessary I suppose!

And yes, if using CLU you get addicted to it!

 

Very important when you are married to a wife who does not accept indoor temperature below 20 degrees and preferably will have it at 25 . And who is ready with the matches in front of the stove when the temperature goes down to 20, LOL