Repasting with CooLaboratory Liquid Ultra, any tips before I start?

Stress test for 1 hour

 

4.6Ghz for 1hr, using Profile supplied by Mr.Fox

View attachment 107307
View attachment 107308

 

Impressive, very impressive :thumbsup:

My finger is getting itchy... I want to click and order some now.

Did you apply some on your GPU too?

 

hehe bro TBoneSan just get some, my constant mental jawdropping made me click on the BUY button earlier already, nothing can go wrong

 

I'm getting excited about testing this stuff.

@[Nikos] - those results are superior. I agree with TBoneSan... if you continue seeing temps that good a month from now, everyone needs to go buy some Liquid Ultra to replace whatever they are using. I would love to see what your 3DMark Vantage Tests 1 & 2 at 4.6 GHz max temps looks like.

@ht_addict - kudos to you also, as you showed us your temps a while back and I see they are still holding up well. I am very impressed by how cool things are still running after all this time. The 4.6GHz 1 hour XTU test was impressive, especially for how old your Liquid Ultra application is. With any severe CPU stress 4 or 5 minutes is plenty long enough to find out how cool the machine will run in normal use, benching or gaming. XTU is far more brutal than almost anything else you would do day-to-day. Any longer than that is probably just extra wear and tear your CPU will live longer without. I would like to see what your Vantage Tests 1 & 2 at 4.6 GHz max temps looks like in addition to [Nikos].

 

Is there a way to remove it completely or does it corrode and stain the surfaces as bad as they say it does ?

Sent from my SM-N900T using Tapatalk

 

Thanks. Both GPU's have been coated as well

 

Comes of the CPU and GPU die easily. Just watch that you don't cause it to roll into a ball and end up somewhere else. As for the heatsink, you can either just wipe it off and re-paste. Or you have to use a little bit of fine grit sand paper and remove. My thoughts and I believe Mr. Fox mentioned this. If you just do a wipe down of the heatsink, what residue is left embedded in the copper of the heatsink is a good thing as its actually filling in the small grooves and imperfections. But only time would tell with continuous repasting.

 

Yes, and I think Coollaboratory mentions something to that effect, too.

 

Here are my Vantage results at 4.6, and 4.4 vs 4.6 . What a difference in temp you get with an extra 200mhz speed bump. Think I'll just run at 4.4 for my daily gaming.

View attachment 107313

View attachment 107314

 

To me, this is by far the most impressive example you have posted, ht_addict. I wondered why you were always staying in the 4.3-4.4GHz range all the time.

For running Vantage physicis tests at 4.6GHz on nothing but normal room air, those temps are excellent. And, look at your CPU power... 98W @ 95°C for the amount of time those Vantage tests run is very nice. I can live with 95°C on this because you know it's going to get lots better than that benching with AC cooling. Big jump in temps? Yes... but also a big jump (around 1800-2000 3DMarks) in the Physics score, so totally worth it. The 32K CPU score is extra sweet because it is running on ambient room temps. You're still short of thermal shutdown by about 10°C so there is still room to push harder. And, more overclocking headroom (rather than lower normal use temps) is exactly what I'm looking for. Lower normal use operating temps will be a nice secondary bonus as a byproduct.

This is 4.7GHz on AC with a fresh IC Diamond application...

 

Wow... That score.

 

More stress for the NT-H1... it's starting to degrade a bit, as expected.

 

My first run in the morning 3DMark Vantage CPU's test:
4.6Ghz
I think my results are a little bit lower that it should be, maybe some driver's issues or something don't know how to improve this value.
Maybe it has to do with the chipset driver... I don't really know...
If you have any clue please I will try it


PS: Now my house is colder than at night so the temps are better. Anyway this thermal paste it's awesome.

 

Those temps are so low! You should just shoot for 5ghz now

 

Thanks. What were your settings in ITU for 4.7Ghz

 

Mr.Fox I would like to try your 4.5Ghz profile if you have one, to test to achieve the best stability possible.

PS: I'm going to test the thermal paste with modded vbios for my GPU's I would like to know your testimonial, What is the best vbios for just one power supply?

Thank you

 

3DMark Vantage (Personal Record):
4.5Ghz
SLI GTX680M @ +300/+700

 

3DMark 2013 (Fire Strike):
4.6Ghz
SLI GTX680 @ +347/+700

 

4.3 XTU Bench test. I had to increase the flex and pri plane a little to get through. Standard fan tables on a 1 week old paste that has been mildly abused



Core max temperature hit 91

 

Something holding you back there, TBone? I was expecting 900's with those clocks.. Temps are healthy though, long as you didn't do that bench in the snow.

 

Not too sure to be honest. I've just replicated Mr Foxes control settings from pg. 5 apart from Throttlestop.
I'm not sure if 4.3 is enough to land me in the 9xx's ever. I did a run a while back at 4.5 which landed me in the early 9's here - maybe you could have a peek at it and see if doesn't look right

 

what about this? Coollaboratory Online-Shop

non-ageing and it looks very safe to use, anybody willing to pull the trigger on this one too?

I wonder how these compare to Liquid Ultra paste?

 

Review says it doesn't perform as well as the Liquid Pro or Ultra.

Results: Air Cooler; High Mounting Pressure - Thermal Paste Comparison, Part Two: 39 Products Get Tested

 

It is interesting that so many reviews exclude IC Diamond. I don't really understand that.

I'm glad that I tested Noctua NT-H1. I can now recommend not using it based on personal experience. It does a nice job at first, about the same as IC Diamond, but durability is poor compared to IC Diamond. After an intense workout it has already failed compared to my month old IC Diamond application.

Compare this with what I posted for a 30-day application of IC Diamond running the same ThrottleStop 1024M test (or compare it to TBoneSan's XTU run on 1 week old IC Diamond) and it's clear that IC Diamond spanks NT-H1 pretty bad.



The Liquid Ultra should be delivered by USPS today. Looks like it is arriving at a good time, because I'm done with NT-H1 now. I'll save it for repasting my ordinary computers. Should be fine for that. I was going to test it on the Alienware 18 as well, but I do not see any point in bothering with it now. It seems obvious that doing so at this point would be a waste of time and thermal paste, especially considering how much hotter that poor little critter runs compared to my M18xR2.

Final Verdict on NT-H1: [ LINK]

 

last monday i repasted with liquid ultra as well. i did the cpu and a day later as well both of the gpus (4930mx / 2x 780m on AW18).

now, when i do a 3DMark11 run running the cpu @4.0GHz (@1.120V) and the 780ms at +185 / +200 i get the following max temps:
cpu:
81
82
84
83
while before i hit the mid 90s

gpus (and this is funny):
gpu 1: 82
gpu 2: 87

with those settings i had around 14.8k - around 15k overall in 3DMark11 before repasting.
now i get to around 14.7k and struggle to get any higher.

what i noticed is, that in the lower temps, gpu 1 is either the same temp as gpu2, or lower - so normal behavior. but as soon as i do something really heavy like a gfx benchmark, gpu 2 can get up to 6 degrees hotter than gpu 1.
i can't explain why, i did the repasting on both gpus exactly the same way.

what is more puzzling to me, is why i have problems to reach a score i reached before even that i have lower temps..

 

Well, take in to account all the background programs and processes may be idling/runing/antivirus updating scanning etc etc, maybe something happened you've updated drivers after repaste etc, many things will count on your score. Maybe now when CPU doesnt throttle (if it doesnt @ such great temp) maybe your GPU's are starving of power? Many things will count...

 

don't shoot me for this...but i just remembered that i updated the nvidia driver to the newest version while drunk on the weekend... =P *shame*..

 

HAHA ^^^^ Just revert back to previous version and re-run the benchmarks and please post the result

 

just did, but same results, i just can't get over 14.7k. even to hit 14.7k is a struggle. oh well, still happy with the better temps
Phsyix Score bounces from 9.9K to 10.2K without hints of why, same settings 9.9k...next run same settings 10.2k =)
i don't feel like repasting to check if the gpu 2 warmer than gpu 1 under extreme load will change afterwards... this time i will stick to never touch a running system haha

ps: the highest i was able to reach so far is 14.8k.

 

sometimes reverting is not as simple, you may want to try do alternative clean drivers re-installation:

1# uninstall nvidia drivers from control panel in full

2# download driver-sweeper and select to search for nvidia graphics drivers to be removed

3# once done remove them, reboot your laptop and open [device manager], see if it states "nvidia graphics cards" or "standard graphics adapter" in display adapters section.

4# If its nvidia adapter specified -> right click and select uninstall + select remove driver, and do the same on secondary adapter.

5# Reboot the machine and proceed with desired driver fresh reinstallation selecting extra option "Perform Clean Install" in nvidia driver setup wizard.

6# Report back to us how did you get on with fresh driver reinstall and benchmarks

 

will do so when i get home and then report back thx a lot

 

The latest drivers had a positive effect on Fire Strike and a negative effect on 3DMark11 and Vantage for me. I went back to 331.82 because Fire Strike is the least valued benchmark benchmark of the three, IMHO. It is mostly a GPU test and it does not seem to be CPU dependent enough to suit me. 3DMark11 and Vantage are a much better test of overall system performance and place more weight in the overall score based on how the CPU performs.

 

i think my installation is about to die... -.- i uninstalled the drivers, cleaned with driver sweeper, then installed the 331.82 driver but i can't start the nvidia control panel. after i installed the whole driver package i can start it once, i.e. set SLI to activated, restart, then it will not start again (the nvidia c.p.). uninstalled and reinstalled around 5 times, no success -.- you can't even imagine how i don't have any motivation to reinstall windows..just did it like two weeks ago xD ;-(

 

why not use nvidia inspector for all the settings adjustments? I know one tip: reinstall nvidia drivers package now without removing them and before you reboot your laptop, nv control panel will start

btw, had same issue, using win7 64 pro, somehow problem went away also by itself, weeks later

 

i didn't even know that it's possible to set sli from within nvidia inspector ;-) so far i just used it for oc... haha
well i'll see what a bench brings, then if it won't fail i will leave it and see if it's ok like this ;-)

 

While initial results looked promising, Noctua NT-H1 ultimately ended in failure after less than a week of heavy use. I cannot recommend this paste for CPU overclockers at all. It cannot hold a candle to IC Diamond due to poor durability. I think it might do fine for GPUs and may work fine for folks that have Intel CPUs running at "ordinary" clock speeds. Check out what a lousy job it was doing with Darksiders II late in the day before I applied Liquid Ultra... kind of pathetic.

The one thing I can say good about it is that NT-H1 was super easy to clean up. It was a lot like wiping toothpaste off of a slick surface. Wish they were all that easy, but that might explain why it does not hold up well in harsh conditions.



Anyhow, Liquid Ultra was applied. See updated results in post #47. All of the images from testing are arranged in the same order to make it easy to compare the temperatures in tests among all three pastes.

I will put it through the same kind of brutal tests as I did Noctua NT-H1 and IC Diamond and report how it is holding up in another four or five days (Saturday or Sunday).

 

i think concerning gpu 2 running warmer than gpu 1 while under heavy load i noticed, that sometimes gpu 2 has 99% load while gpu 1 has 98%. it's just 1% but maybe this is just enough ;-)

 

This is pretty amazing... 45x4 with nothing but fans running Liquid Ultra


click to enlarge in a new browser window

 

Wow, great temps. I really hope this TIM is durable. It would be cool having 4.5 as a 24/7 profile. Crazy stuff.

 

Yeah, look at this one...


click to enlarge in a new browser window

I'm looking forward to benching this on AC cooling this weekend.

 

so far looks nothing but absolutely amazing to me

@TBoneSan - I was doing some random read about it on web - people "complain" about this TIM to dry out and start to do a bad job 6-9 months down the line... I found this to be a very good "complain" !

Of course application to application and terms of use will always vary and results will differ but I cant imagine this TIM turn bad in a month or two to be honest.

I was going to be happy if I can re-apply it every 30days, but if its going to last me even longer - that's a win/win for me

PS: my PhobyaHeGrease Extreme (ceramic based TIM I used last time to re-paste everything) I feel like it turned me down day by day, it was going good for first two-three weeks and now I see daily stock clocks peak mid 90's on just heavy desktop use, no rendering , just massive multitasking and background tasks (RAM load averages 10-20GB in use), but CPU @ stock clocks..... I will be doing major repaste on everything once this LiquidUltra comes in trough my door

 

Looks good. I'm worried about the removal though. I've heard you need all kinda of sand paper and stuff

Sent from my SM-N900T using Tapatalk

 

^ wrong

all you need is isopropyl alcohol cleaning agent and soft fibre disposable cloth, wet the cloth and wipe it off gently. You can see demonstration video on their website how they do it. Easier than seeing woman (if you got one) removing her make-up of her face every night

 

I think DanDan might be referring to is cleaning the heatsinks. But from what I'm gathering it's not such a big deal if you're going to repaste with the same Liquid metal again.

Does anyone recommend a particular brush to spread it with? I'm thinking any fine brush with softer bristles would be the way to go yeah ..

 

if youre buying retail pack - it should include brush and cleaning material in the kit. Other than that I recommend reusing the supplied brush by cleaning it with isopropyl alcohol after application.

I know some places sell the syringe of TIM only as a bulk package but price doesnt always appear to be lower

 

I used the brush... kit includes two of them. I cleaned the one I used with alcohol.

I "painted" a thin layer on the die and heat sink contact surfaces. It takes hardly any Liquid Ultra. For both surfaces I used approximately this much > O (yes, about the size of an uppercase letter O). A little bit goes a LONG WAY.

For those that might be interested, here is how the crappy Noctua NT-H1 paste looked after removing the heat sink. Obviously, the stuff is way too thin and "pump-out" is an issue.


(click to enlarge in a new browser window)

 

Nice, so no permanent staining with this ?

Sent from my SM-N900T using Tapatalk

 

I don't know, but at this point I don't care about that. As long as it works well and lasts at least as long as IC Diamond does, then I am OK with it. Stains in the copper surface of the heat sink should not matter as long as functionality is not impaired.

At this point I do not intend to repaste the video cards with Liquid Ultra. I think I will stick with IC Diamond for GPU paste. Even when heavily overclocked and overvolted, they don't get anywhere nearly as hot as an overclocked XM CPU.

---

Edit: Nice video rendering temps at 4.3GHz... never exceeded 83°C.



<iframe width='640' height="360" src="//www.youtube.com/embed/RISw1E0MLiA?rel=0" frameborder='0' allowfullscreen=""></iframe>

 

Good runs! You know you can remove a lot of the extraneous information in Hwinfo64 right? Here is what mine looks like, much easier to read, all the critical data: