Clevo Overclocker's Lounge

That looks kind of cyan... I kind of want.

 

BTW, right now I got IC Diamond 4 or something, and it heats up to 100C on 4.1 GHz, is IC diamond usually better?

 

ICD7 and ICD21 are just the amount of paste in the tube as far as I know. Same product.

Never heard of ICD 4, though.

 

The thing with IC Diamond is it is extremely durable and works better than most pastes short-term and especially long-term, and if you have a sloppy fitting heat sink it's the only thing that works well at all. IC Diamond can be used to fill some abnormally sloppy gaps and still work fairly well where anything else will fail miserably. Clevo resellers and Sager dealers are very smart for using it considering how sloppy many Clevo laptop heat sinks are fitting.

You really should consider using CLU at least between the die and IHS. Unless you're totally reckless and couldn't be careful if your life depended on it, there is nothing to fear... especially between the die and IHS on a 6700K. There are no FIVR components to get shorted out like a 4790K. There is nothing in close proximity to the die except silicon.

Yes, the Plexiglas "lens" makes it look kind of cyan. It does look really nice.

Yes, identical in every respect except for the volume of paste in the syringe.

 

If you apply something like nail polish or electrical tape to any exposed conductors under the IHS it's almost impossible to mess anything up. You can then either seal the IHS down with silicone or fill the gap with a traditional paste so there's zero risk of any leakage.

A little bit more blue than I'd like, but maybe it's the camera. Still a ton better than the defaults.

 

Finally working back into 4.9GHz... little by little.

 

Sorry guys, been really busy the last week and now i'm in Las Vegas for CES. It is seriously packed out here. And I have seen a few things of interest to us all in terms of cooling. Trying to see if it can be done for a laptop. They said they will be looking into that.

Thought I would post this to say i'm still able to validate 5.0 ghz on first gen Prema bios .07

 

Holy voltage Fox-man, did it always need that much?

Can I come next time? I promise I won't get thrown out... until the last day... for asking too many questions... and generally being me.

 

So far, most seem to like answering questions.

 

http://www.3dmark.com/3dm11/11883791

Yes, and the problem was after I installed the previous version of CCC, I could not give it enough and it kept dropping too low under load. But, if you are looking at idle voltage, don't. Load voltage is barely enough at 1.360V. It needs to be around 1.385V under full load at 4.9GHz and that's not very high. See the example below.

 

Great that it was fixed. But what was it that made the steak?

BTW http://www.guru3d.com/news-story/am...d-running-3-6-ghz-base-and-3-9-ghz-turbo.html

 

ha, nice choice of color there, ive also been using cyan for quite a while now

 

It is actually white, but the Plexiglas lenses in the slits give it a cyan cast. I like it.

I have no idea. The only things that changed were the introduction of the previous version of CCC and the EC v08g. Why it has been difficult to undo the damage is unknown. I had saved photos of BIOS and ThrottleStop settings that suddenly no longer worked that I had been using for benching with excellent consistent and repeatable results, and I still do not understand why... probably never will. And, they still do not work. I've had to start over completely from scratch. Getting to 4.8GHz has been really easy. Getting 5.0GHz validation and benching stability at 4.9GHz has been very difficult because of load voltage drops that I was not having before. They are small drops, but just enough to upset the apple cart.

What is weird is passing Cinebench and standalone 3DMark Fire Strike and 3DMark 11 Physics tests with flying colors, but during the full benchmark having it run like a banshee only to unexpectedly freeze in the Physics test. The part I don't understand why it passes when run as an independent Physics tests, but when running the full benchmark it crashes with 0x124 and 0x101 (voltage) errors. The Physics tests workload is the same. I am going to try with GPUs stock and SLI disabled and see if that changes anything. Maybe the power demands of the GPUs in tests leading up to the Physics tests have depleted reserves or something. Maybe switching ThrottleStop profiles to 4.9GHz only at the start of the Physics tests will work.

If I cannot figure it out, I'll see if Brother @Johnksss@iBUYPOWER can help me understand what is happening and how to circumvent the problem. It could be something I am forgetting, but I don't know what that might be. But, I'm getting really close to where I was before with very different settings than what used to work more or less flawlessly. I am only able to do it again now while using the current @Prema BIOS mod. It's probably some kind of wiener-boy power-saving "efficiency" cancer feces causing it.

 

I just finished applying Coollaboratory Liquid Ultra. Took me a full hour; I was working from 1:10 am local time to 2:10 exactly. I ran a P95 Small FFTs test before, and temperatures shot right up to 90°C, and stayed there. I don't have a 'before' shot (ah screw me, I was too over-enthusiastic to see lowered temps), and here's an 'after' shot of temps while running the same thing for 5+ mins:


The max temperature was 95°C, but the slow climb was the nicest thing. Is there a curing time for CLU? I'm thinking of letting either Witcher 3 or X-Plane 10 run overnight if so...

Edit: I also noticed that temperatures climb much less slowly as compared to what I had (Gelid GC Extreme from my reseller, I think). But somehow the max temperatures under a very very intense load isn't that much better... I'm slightly put off when I see miracles on Amazon where delidded CPUs get 20°C lower temperatures.

 

Well done. + rep
No curing time for Liquid metal. But re-test all test you have done with previous thermal paste. If you remember previous temps that would be nice to share.

 

Thanks!

Eheheh, I really was too over-enthusiastic, forgot to do thorough testing before and after.

I can be quite sure though—P95 small FFTs used to give an immediate rise to 90-95°C, and hovered around there. Now, temperatures rise very gradually to 95°C.

As I mentioned in my edit, I was expecting a lower maximum temp as well, maybe around 75-80°C based on reports from Amazon reviews and this forum. I guess the Core i7-4710MQ really is too much for the W230SS... How does @Prema do it?

 

Okay, just got me a 7700 k, but wont get to play with it till i get home.

Sent from my SM-G925P using Tapatalk

 

Really simple. Benchmarks don't last long, and don't hammer you with AVX like P95 does.

Liquid Ultra can lower maximum temps, sure. If it's within the ability of your cooling system to do so. If it is not, then your observed behaviour where your temperatures rose slowly instead of instantly hit the limit shows that it's working properly. But maybe your cooler just cannot handle it. Is the laptop propped up? Max fans used? If so to both, then maybe try other non-AVX benchmarks and see how it works.

 

Small FFTs for me maxes out at about 87-89C , when it's on tops of my U3. This is with DIE->CLU->IHS->CLU->SHIM->CLU->HEATSINK.
Without the U3, tops out at about 92C.
Ambient temperature at about 25C.

But even with the worst of the worst day to day CPU loads you will not even come close to the small FFT load.
Try AIDA64, That will be much more realistic.

 

Fair enough. I used TSBench, and temps hover at around 65°C, compared to about 80 previously.

Indeed, yes to both. As I mentioned earlier—the W230SS cooling system really would work much better for the i5-4510M instead of the i7s, although @Prema blasted his 4700MQ to, what was it, 3.9 GHz, and @t456 has a 4930 or 4920XM in his one?

Nevertheless, temperatures in Witcher 3 at the heart of Novigrad with Background Characters turned to Ultra keeps temps both on the CPU and GPU within 70°C. Fans were on auto. They used to rise, unshackled, to 85°C and above. All in all, under low–moderate loads, I have lower temperatures by about 10-15°C.

I really should've tested temperatures properly beforehand...

 

Eh, it's fine. Now you know for sure that you see a benefit, which is great. AVX is just too much for most notebooks to handle, I suppose. Also, fun story, I typed the entire last message and this one with my eyes closed. For no... real... reason whatsoever. xD

 

Just checking, AVX is an instruction set, correct? Why does using it put out so much heat even if power draw is relatively low?

How the hell???

I'm going to attempt an ffmpeg transcode first, see how it works. But this is for tomorrow. It's already 3.40am here, and I am dead tired.

 

FIVR for some reason gets owned by AVX. Haswell and Broadwell get abused left right center up down around the world and in the toilet bowl hole. SB/IB/Skylake/Kabylake/etc won't.

how? It's actually rather easy xD. For me, anyway. I actually type faster with my eyes closed than with them open, and I don't know when I acquired this skill or how I cultivated it.

Good luck!

 

So... P775 GTX 1080 stays at 70C while their GTX 1080 stays at 75C ....

Also, they said nothing about thermals in the CPU.

Not sure why, but it feels like thermals using slower fans would probably result in not so good temps, but much better sound noise levels.

 

I don't have a 4700MQ, it's a 4702MQ:

http://hwbot.org/submission/3168619_

 

A 4930MX atm (4900MQ ES before). The W230S* can cool it just fine, but it is all liquid metal these days.

Bought an 80g bottle of the stuff a while ago, which is equivalent to 80 syringes of CLU (at 1/9th the cost), so whenever someone's system passes my hands for one reason or another then they get a CLU treatment. Could do even better when you're a reseller; for 250g you'd have more than enough for over a thousand machines, so material cost is a dime each. Really ... this kind of bulk makes you wonder whether it'd do to substitute for water in a water-cooled desktop (using a far stronger pump, of course). Its density is about 6.5x that of water, which'd make it behave (and perform) like fluid copper.

 

That type of project sounds rather evil, but really want to see it done now.

Like really curious to see it put in place!!!

 

Sweet! Congrats. I can hardly wait to see how high it overclocks and benches. I'll be especially keen to find out if the rumor that Kaby Lake requires Windoze OS X is actually true or just one more example of the baldfaced lies, deception and manipulation crap that the Redmond Nazis are becoming famous for.

 

I can also pick one up right now, but @Prema San is busy , and I don't want to mess with his files trying to insert micro codes.
Haha
Really want to see how well it performs in the DM1.

 

I really want one, but it will probably be at least 6 months until I can work that extra purchase into my budget. So, until then I will have to be an observer.

 

Something was different with new bios....

Yeah, that's bs. I was at the hwbot OC competition and they all had 7700K's and benching windows xp /windows 7. But then they also had unlocked boards as well. They said I should have gotten there earlier and joined in on the fun. I will next year for sure. And they also have a new deliding tool kit. It seems to come with everything to delid and reseal the chip.

 

From the comment section. Probably tested on a desktop.
http://www.tomshardware.com/reviews...00k-i5-7600,4870.html#comment-rsp_en_19098433


Ps. This is what they do in the AW section http://forum.notebookreview.com/thr...3-owners-lounge.797458/page-209#post-10428106

Maybe they need it? I mean disable turbo boost!! http://forum.notebookreview.com/thr...3-owners-lounge.797458/page-208#post-10427426

 

http://valid.x86.fr/4fh63v

 

Haswell: 1.55v
Skylake: 1.385v

Well then...

 

check out the current stats on 7700K CPUs tested by silicon lottery (as of Jan 5th):

100 % of tested CPUs reached 4.8 Ghz stable OC (1.376V max.)
91 % did 4.9 Ghz (1.392V max.)
62 % breached the magical 5.0 Ghz barrier stable (1.408V max.)
24 % still did not give up and went to 5.1 Ghz stable (1.424V max.)
4 % cream of the crop can do 5.2 Ghz stable (1.440V max.)

now those are pretty damn sweet numbers, especially considering theyre just for the very first batch!

 

http://arstechnica.com/gadgets/2008...quid-metal-cpu-cooler-terminator-jokes-ensue/

http://www.guru3d.com/articles-pages/danamics-lmx-superleggera-review,1.html

I wasn't able to find much about this, and it's mostly 6 to 8 years old. Except for this DIY project with no apparent progress outside planning stage.

I believe it's one of those things that seems cool but happen to be too impractical, expensive and/or with no real improvement than current tech, and become forgotten soon. Otherwise in 8 years we would have seen more.

 

How big was the first batch?

 

i meant the wirst wave of production samples from intel, since we only recently started to see Retail CPUs of the 7700K SiliconLottery didnt specify their sample size, but it should be quite large in order to be able and sell sufficient quantities in their online shop...

if we say that 4 % has to equal at least one CPU, the sample size can be calculated to a minimum of 25. id wager it would be larger than that!

 

Either I got very lucky or those are very strange results...

 

Run a longer stress section is different than short benchmarks. You need a bit higher Vcore for 24/7 stability. But around 0.048V lower vcore vs. previous [email protected] SL chips is not very impressive.

Edit. Yeah, Those vcore numbers is higher than from most of the short pre-reviews of 7700K http://forum.notebookreview.com/threads/kaby-lake-i7-7700k-findings-not-for-fcbga-aka-bga.799156/

 

Well I went ahead and ordered that last 5.1 he had listed on the site. I know its early and was a 90 dollar premium over buying one on the shelf but what the heck. I looked at the address and noticed the operations are set up about 20 miles from my house. I had already posted in the 870DM3 lounge but let me know what (free) testing you all want done so there can be additional testing comparison here.

 

The more I look at this turd, the more I hate it. Ignoring the BGA part momentarily, that is an ugly sucker. Linus's video really gave us a better look at how tacky it is. I don't have think they could have made it look more like a Hasbro sci-fi child's toy had they tried. Really, the only redeeming qualities it has are its massive dimensions, the display and keyboard. The flip-over 10-key/touchpad has an element of geek novelty to it, but functional limitations since you cannot use both at the same time. The P870DM3 is superior in every way that matters most for half as much money.

 

1) The only thing I love about it is the large screen and keyboard. I've always wanted a taptop with at least an 18.4" screen like the AW18 and the bigger, the better. think of it as a portable desktop

2) I never use the trackpad so the ability to always have the number pad and not see the damn trackpad which makes me feel like a disabled human is a plus point

3) no Prema BIOS and the name ACER just doesn't excite me

4) The price.........ughhh..... I didn't win the lottery yet..

5) For that price, you could get to supped up P870DM3s with Kaby Flakes that would run circles around this

 

dont forget, those VCore numbers are MAX. values. it stands to reason that most of the quoted CPUs reached those clocks with much lower VCore

@Prema of course youre lucky, i wouldnt expect anything else, karma and all

 

And I will laugh until the moon if this big smartphone deliver worse performance than the other BGABOOK's with the same retarded innards. As I said earlier... Already destroyed before it came on the drawing board!! Only intended as an advertising poster for Acer's equally crippled Jokebook brothers-sisters. You can't change it with a nice bigger curved screen. Born to be Crippled!!

 

The guys love these new chip. They say it is better than the 6700K. So that's what I'm going to go by. Hope I win the lottery, but if not. I know where to get one.

 

It boots 5Ghz on stock voltage, so we should see some 5.5Ghz results from you here soon...

 

If iGPU + a bit higher clock had been the only thing that was improved with this revision, would 7700K been a failure. A huge failure