http://archive.benchmarkreviews.com/ind ... mitstart=1

I am just reading the above link. If carbon based TIM is better, and assuming we are comparing the end carbon based TIM vs. top end Ceramic based, so say we have:

Arctic Cooling MX-4
TIM Consultants TC Grease 0098
Shin Etsu MicroSi G751
Cooler master Fusion 400

Now, they all seems to be at the top quality level, if so, would you pick MX-4 as it's carbon based vs. the others ceramic based TIM?

He tested 80 different compounds. If you toss out the worst 5, then the temperature difference between the best and the 75th ranked is 2.45C difference over a ~40C rise in temp. Don't get me started on test repeatability... The answer is it just doesn't matter unless you are a crazed over-volter. Just pick one that:
- is easy to apply/remove
- doesn't cost a fortune

I like the electrically non-conductive pastes (like Arctic Silver Ceramique) so I never have to worry if/when a string of it accidently goes somewhere it shouldn't.

http://www.hardwaresecrets.com/article/ ... 012/1490/5

but take a look at this test from hardware secrets, the range is from 47 to 63 deg. C w/ various different TIM

They tested things like Chocolate, pink lipstick, toothpaste, and butter. Surprisingly some really odd things done alright, i mean who would have thought that cream cheese would be only 37c rise from ambient

EDIT: Damn emotes i miss you..

Olin Coles, the author of the amazingly well done Benchmark Reviews Study, had this to say about choosing TIM: "My conclusion clearly states my future intentions. [Which is not to do any more TIM reviews.] It doesn't matter how many new TIMs come to market, if you properly prepare the surfaces and ensure great mounting/clamping force, they'll all perform nearly the same."

Happy Hopping: Read more carefully; the 63C high you mention from the Hardware Secrets roundups is WITHOUT ANY TIM. The real range they found is 32-40C, based on a quick scan, and they state their error margin as 2C, which means anything less than that is meaninless (ie, the relevant temp variance is thus merely 6C). I also don't think they were quite as meticulous in following manufacturers' application recommendations as closely as Olin Coles, which makes their results less reliable.

Anyway, in essence, I concur 100% with CA_Steve.

AS Ceramique is my TIM of choice, but I would probably get some of the very high end stuff if I were to do advanced things like delidding CPUs.

Personally i just like to buy none conductive and none capacitive compunds, for the same reasons as CA_Steve, my preferred due to the before and cheap is Artic MX-4. Lately i been using IC Diamiond 7, and good results, but barely any difference over MX4.

I was leaning on MX-4, as I am running out of AS5. But I am confuse about this:

From this link, it is clear that Silver has a much higher conductivity than MX-4.

http://wccftech.com/review/thermal-grease-shootout/

now from this link:

http://forum.hwbot.org/showthread.php?t=71658&page=2

AS5 is at 1134 w/ m *k

MX 4 is at 8.5 W/m*k

So the conductivity is a lot less on MX4, and how does it better than AS5?

Mike, I am well aware line 1 is w/o TIM. But I was looking at column 2, under Core temperature:

igoring line 1 and item like chocolate, we are talking about a core temperature of 47 deg. C (for e.g., Prolimatech PK1) to 63 deg. C. (ThermalRight The Chill Factor) i.e., some TIM can do a core temperature of 47 deg. C, and yet some other can only do 63 deg. C. That's a 16 deg. C between different product, and that's quite serious.

I much prefer my CPU at 47 deg. C vs. 63 deg. C. Anyhoo, do you have any insight to my question, 1 post up about silver conductivity vs. others?

You can't just look at core temperature in isolation, as the room temperature varied. So if you like 47°C, I'm afraid you'll have to move somewhere with bad heating.

Prolimatech PK1 = 34c above ambient
ThermalRight The Chill Factor = 37c above ambient

As Cistron said you have to look at the rise from ambient (room temprature), i had trouble finding anything more than 10c

Okay, I miss that. I didn't realize they have room temp. that low. So in the end, it's a handful of degree of temperature only. From the low to the high, about 6 deg. C

6 deg. C is not that serious, but if my CPU can drop 6 deg. C, it would still be a very good thing.

Why?

Lower CPU temps are always good surely? Whether it's worth bothering with, depends on how much effort and money it takes.

The liquid metal kits void your CPU warranty, cost a fair bit, and take extra effort to apply. Hardly worth it for most of us when decent thermal paste is only a few degrees behind, but certainly worth it when going for overclocking records.

I would never do liquid metal. What a headache that would be. MX-4 I would do. Does any1 knows how many Xeon CPU can 4 gram do?

Now I'm starting to wonder how Mercury would fare ...

simply not true. if its cooled quietly and still below spec temp, and not throttling then there is no gain in lowering temps.

Sometimes indirectly allows for fans to be turned down a bit though.

IMO there is a gain, in the form of increased thermal headroom for anything from overclocking, to seesawing ambient temps, to dust buildup over time and gradually increasing system temps. It may also allow you to run your fans that little bit quieter while maintaining the same target temperature. That's surely worthwhile?

Still, it's a moot point really, as the difference between the pastes is pretty negligible. But if the price difference is tiny (per gram of thermal compound, not just per tube), then I think it's worth it to choose the best value paste.

all im saying is; if your cpu/gpu are at your preferred clock speed(s) and under spec temps, with your fan(s) quiet enough to not hear over your ambient noise level, then 6 degrees less give no benefit. i see it all the time, -some one is happy with the noise of their rig at 65*C when maxed out in prime 95/furmark ect. , but for some reason they still want to lower cpu temps. (and often their real usage would never reach that kind of stress anyway.. )

BTW, if you're really worried about a couple of degrees, you should probably look into lapping your CPU heat spreader or removing it all together.

Ah yeah I see what you mean. You're talking about after you're built the PC, it not being worth opening up the rig, removing the cooler and re-pasting just for a few degrees.

My view is from the initial purchase of the aftermarket CPU cooler, when you're doing research and deciding what thermal compound to buy (if any). If two pastes are the same price or close enough, and one will give you 2-3 degrees lower temps on average, then of course you'd get the better paste. No extra effort or cost, and something small to gain.

I have used 2 TIM's in my time. Years ago I always used arctic silver 5. But arctic silver 5 has been out a long time so I got a new paste and did a simple test.http://www.newegg.com/Product/Product.aspx?Item=N82E16835426020 - Gelid GC-Extreme. Heard good things about it so I decided why not give it a go. On a Sandy bridge 2600k system and a noctua u12s the gelid gc extreme was 4!!! degrees!!! cooler than AS5. I was actually able to lower my fan 200 RPM and get the same temp as AS5. I can 100% confidently sat Gelid GC extreme is in the top 10% of TIM's and it's 12.99 for a 3.5g syringe, not the cheapest but isn't 13 dollars worth it to be able to turn your cpu cooling fan down a bit? I've heard coollaboratory liquid metal is supposed to be the ultimate best but just look around at all the problems people have had with it. It's not viable for long term stability.

Just giving my personal 2 cents on what I've experienced with the 2 TIM's I've used. I don't think anyone here would be disappointed with gelid gc extreme.

I'm well into my second 22g syringe of Arctic Silver Ceramique. A 25g syringe is $7.99 @ Newegg and AS Ceramique also outperforms AS5. Might not be quite as good as Gelid GC-Extreme but a whole lot cheaper.

How well or how badly TIM is applied probably has more influence than the actual TIM, AS5 tends to be very viscous so can be difficult to get a good application imho.

With AS5 i use a brand new razor blade to squeegee the thinnest coat possible. works way better than a credit card.

Highly recommend gelid gc extreme. My personal experience is it dropped a sandy bridge system 3C cooler than Arctic silver 5.

When it comes to applying AS5 it depends on the CPU you are using on how best to apply it.
http://www.arcticsilver.com/intel_application_method.html#
I have seen temps vary by 6-7c from a poor to good application i guess its why some people find other TIM's to be better, AS5 can be a real PIA to get right. It took me three tries a load of reading and watching YouTube clips to eventually get a result i was happy with. There is a lot to be said for TIM's with better spread and easier application than AS5.

but if the CPU temperature can be lower by 6 deg., that CPU last longer

I have an Athlon XP CPU has happily lived for 8 years at ~52C in idle(target temp in speedfan), survived an overtemp shutdown(a bit of roasted electronics smell included) and that CPU has lower tolerances than Intel's.

Removing the IHS as pointed out would be the approach if you're serious about dropping temps, that said, there are risks involved and you will void the warranty.