Clevo Overclocker's Lounge

Let's up this and imagine the CPU is a new A series that destroys Intel and AMD with an in house GPU that also kills Nvidia and AMD. I believe the world will be a better place with BGA being history and Nvidia releasing powerful firmwares with their GPUs to counter the monster.

 

I like @Mr. Fox 's earlier suggestion. Try it, perfect it's application and use, then you can speak from a position of authority on it's performance

Then try Noctua's NT-H1, that stuff is the "bee's knee's"

 

Very true!! Fox is smart.

 

What is The Best Thermal Paste 2017? – Thermal Paste Comparison
https://play3r.net/reviews/cooling/...son-2017-what-is-the-best-thermal-paste-2017/
What's The Best Thermal Paste? | Thermal Paste Comparison 2017

 

Anyone know of a browser plugin that automatically repeals and replaces all pictures of that person?

 

Exactly. One single degrees C between Liquid Ultra and coolaboratorys copper paste on 3rd. I mean @bloodhawk or @Mr. Fox have tested the copper paste. Would this results work likewise for cpu in laptops with worse cooling? Nope

 

Under most circumstances this is true, but it is assumed the product you are getting more of is the best product available.

I still think that is the best thing to do. It's not like it is difficult to do. @Phoenix - your experience might be different. I doubt it, but even if your experience proves it is garbage, people love you and trust you. If you tell them you used it and it was junk, just think how many noobs you will save from MX-4 hell. If you don't test it to see how it works, you would only be repeating what others have said. Even if what others have said is accurate, it's not a position of authority based on personal first-hand experience. And, if the opposite turns out to be true then we will know there are exceptions to every rule based on your example or a positive experience.

The list is incomplete without Phobya NanoGrease Extreme. Everything I am seeing puts it right on par with Kryonaut, but a shade better because the viscosity is just a tiny bit heavier. It's really good based on all that I have seen. @Phoenix, if you don't want to use liquid metal, I recommend giving this stuff a shot. You saw the Cinebench temps at 4.7GHz I posted a page or two back?

Really? I must have missed where that was posted. Did @bloodhawk post opinions and results from the Liquid Copper? I had been wanting to try some of that stuff and just haven't gotten around to it.

 

You're welcome!!

 

Making laptops everywhere great again! Gentlemen, we have the technology... to build them better... stronger... faster... don't believe the pop culture lies about BGA being good.

Spoiler: Off Topic

I had some reservations and cast a vote for him based on it being the clear and unmistakable lesser of two evils, but I honestly couldn't be happier now. The fact that the media goes out of their way to misrepresent and mischaracterize him based on their own screwed up agenda makes me like him more and more. We just need the clowns that are resisting him to hush up, go away, and let him do what we the people elected him to do. It sucks that they are all such sore losers, but at least now we see a clear picture of what terrible people they are... will stop at nothing to have their way and act like spoiled children when they do not. You did not see any of us civilized conservative folk raising hell, rioting and acting like retarded children during Obama's 8-year reign of terror and war on the American family. I only post this to demonstrate there is a need for balance and an opposing view to the filth that the talking head minority on TV want us all to believe. If they keep it up a second term will be the inevitable result of their ridiculous behavior. They are slitting their own throats.

Gracias, Amigo!

@Phoenix - Phobya NanoGrease Extreme

 

Is liquid copper a paste? If so I may have to try that.

 

Yes, made by Coollaboratory... but check the link from @Papusan as it wasn't any good based on @bloodhawk testing. You'd likely be more satisfied with the Phobya results.

 

I see. My friend ordered some so I'll take some from him. I've already repasted two of his computers with IC Diamond before so I'm sure he won't mind.

 

Be very careful with the applications. It IS Conductive.

 

And no Thermal Grizzly Conductonaut either

 

@Mr. Fox
There are a lot of reviews now online for paste, I found one I liked, it shows NT-H1 and Arctic Silver against Liquid Metal.

If you want a different mix of pastes tested in a review, you can find it and share it here.

 

U guys posted earlier a pic where different methods for applying liquid metal where mentioned. I always thought there is only one method (spread), so what do they mean by standard or spread method?
@Mr. Fox @hmscott @Papusan

 

It's only one way, bruh!! Or expect buying a new pc or RMA what you have on your COST!! Paint the LM on Die, IHS and heatsink with a paint brush or similar tool. If not expect you will smell fried hardware directly after you turn on your computer or very soon *If* your pc can manage to boot up In Windoze before the hardware being fried!!

 

Yes, @Papusan is correct. Anyone that attempts to apply liquid metal without spreading it with the tools provided (CLU includes a brush and Conductonaut includes a dense cotton swab) and letting it spread itself is simply looking for problems. That does not work. It has to be 'painted' onto the mating surfaces to achieve the appropriate outcome. There should be no excess or puddles on the mating surfaces. This is why liquid metal is not a good option if the heat sink fit is crappy. You cannot add more to plug the gap or compensate for sloppy manufacturing.

 

Yes, due to the viscosity being so low you do need to spread it manually, it could just squeeze out instead if the fit is not flat.

 

See also my older post in this thread. As a warning for all of you who never have used LM before!! Not all YouTube videos out there is useful for obtaining proper info!! THE END RESULTS IF YOO PUT ON TOO MUCH LM! FRIED HARDWARE!!

Exactly. As Yoo know, people can misunderstand info.

 

I am wondering when will we see soldered heatsinks in laptops.

 

Soldering a copper plate to a sensitive die is not easy, note that intel are moving away from this due to the cost and yield impact of it.

 

Thermal paste glue will be the next Can I smell Alienware?

 

Many people used to complain about CPUs like the 4930K and 4960X being soldered and not delid-able. It's unfortunate that Intel uses such lousy paste under the IHS, but I would prefer to have the lousy paste and retain the ability to delid and use liquid metal rather than having it soldered and not be able to take any steps to correct a temperature problem.

 

There are thermal ceramic adhesive pastes, and since BGA can't be easily replaced, it would then follow that permanently affixing the cooling to the CPU / GPU if it made the results more consistent would be "better".

They'd have to design the heatpipe and heatplate paths carefully, as disassembly and removal of all of the cooling hardware from the motherboard would be much more difficult.

The design could allow the connecting parts to be removed, leaving the heatplate's affixed to the CPU / GPU, so less area would be permanently hidden under the plates. Basically a permanent metal "IHS" for the BGA CPU / GPU.

Hmm, that might solve the inconsistent assembly by automating the heatplate "gluing", for a solid and even coverage.

 

Honestly, if the end result was consistent and the functional equivalent to a proper paste job, and it was durable, (guaranteed for a minimum of 5 years with no thermal degradation,) it would make as much sense as soldering the CPU and GPU. If the CPU and GPU are already soldered crap, how would permanently attaching the heat sinks make it any worse? It is already disposable junk, so I fail to see a greater level of downside to that. Of course, those are some mighty big "IF" things to expect them to execute on consistently. Most notebook manufacturers already struggle with inconsistency in one or more areas, and getting that right 100% of the time would be absolutely mission critical.

 

If the CPU / GPU heatplate were treated as a kind of extended IHS - with thermal connectivity for heatpipes, cooling hardware, maybe even affixed before placement / soldering on the motherboard so as to have a controlled jig for alignment, it might actually improve current variations in assembly.

Where's that $100 idea box when you need it?

 

Definitely not for me, but hey... it could work, and there would be plenty of people stupid enough to buy such an abortion. It could turn out out working better than a crappy tri-pod or two-pod piece of trash.

 

A soldered in IHS made of a vapor chamber providing a nice, easy and large contact area. Sadly it would be quote a bit taller.

 

Hence I suggested Alienware as the bell sheep. Can't be any worse than the 3rd flimsy TRIPOD leg Let 3M design a new thermal glue for Azor and we may see it within 3-5 years <The year 2021> Say welcome to the first 0.6-0.8 inches Awbooks without the well known TRIPOD HS design!!
Razer, Aorus and EVGA have to step up now. Or Dellienware will win the flimsy- thin game.

 

They could follow crApple's example and move to unibody design. They could improve upon their filth model by soldering the CPU and GPU die directly to the aluminum chassis for passive cooling and connect the CPU and GPU to the motherboard with ribbon cables, LOL. If your GPU dies you replace the bottom half of chassis and CPU instead of the motherboard and CPU.

 

Hey, if GPUs do go HBM perhaps we can have a socket in replacement instead of a card.

 

That could actually be pretty cool if they have a standard and stick to it for a long time.

 

I would go with that in a heart beat! Just switch out the chip like a CPU. If the interposer is a worry, maybe a good soldered IHS on the graphics chip also.

 

And after you follow Papusan's instructions, PLEASE take Mr Fox's advice and use your FREE INSURANCE by placing light compressible foam with a cutout in the shape of the BGA slug (with some extra space to spare unless you intend to secure it with nail polish remover), so that any LM that does happen to run out gets trapped on the foam instead of turning into a conductive ball of doom as soon as you try to carry the laptop and it gets a hard bump from something. The foam needs to be a NON DENSE light foam--NOT styrofoam, but more like a very light mesh-this type of foam will compress to literally micrometers if it isn't too thick. Do this for free insurance or don't be surprised if your laptop suddenly shuts off with a hiss and spark and never turns on again.

Make sure you either apply high quality electrical tape around any exposed resistors and THEN put the foam on after, or you can paint over the resistors with transparent nail polish remover and let it dry, then put the foam over it. The foam is NOT for being some substitute for tape or nail polish remover over resistors (both polish and tape work to protect resistors)---it's to act as a Dam to prevent any LM that gets off the silicon from getting on the motherboard. And it works.

And do NOT USE LM on crappy Tripod heatsinks or any heatsink with horrible mounting--use Nanogrease Extreme or IC7 for bad fitting heatsinks! Unless you want LM sitting around with purple magic sparks coming out of your motherboard.

 

Is flashing custom vbios still a dream?

Might be a stupid question, but what about bga laptops that have the vbios built in the bios like the Asus G701VIK? What would happen if one dumps the bios, extracts the vbios, modifies it, put the modded vbios back in the bios and flash the new bios? Does Nvidia have a check implemented there too? Would the laptop brick because the system would detect the modded vbios and would not allow it to load?

 

Thought I'd do some benchmarking before reinstallation.

https://www.3dmark.com/fs/13488894
https://www.3dmark.com/3dm11/12349715
https://www.3dmark.com/spy/2300280

 

Yes, I think so. Only hardware programmer from here on out. The goons at NVIDIA are hell bent on trying to control us. I understand AMD is pulling the same stupid stunt with their GPUs now, so we're all screwed. At least they are not discriminating against laptop owners and extending the same discourtesy and disrespect to owners of desktop GPUs.

I have never seen a brick. The flash operation just errors out because the firmware is not signed. NVFlash checks for signature before anything else occurs and will stop if it is not signed.

Hardware programmer is a pain in the butt, but it is safer. Anyone that does much flashing really should own one and take a stock chip dump from everything before starting so recovery is not complicated.

 

Right! I completely forgot that I can use a working card to boot it from the master slot then flash the bricked one in the slave slot, . Thanks for reminding me!

 

i know this is completely utterly the wrong thread, but I JUST BROKE THE MSI POWER LIMIT! Thanks to brother @Phoenix. I'll have to see how to reproduce it.

Just drew 250W from the wall with 0.1% lost from the battery over 30 minutes, no power limit throttling. When I get home I'll have to find exactly what triggered it.

 

If you can replicate it please let us know what the deal is. Excellent news. Maybe the same hex can be used to muzzle an MSI throttle cancer demon on the 16L13.

 

I just got home. I think it's just 2 bytes. If that's true however, it means you should be able to allow a 460W power limit by using the 1070 SLI values. But I changed every byte that was different between mine and Phoenix's EC and nothing worked except the last 2. I did manage to disable my SCM, which was annoying (assuming Phoenix must have SCM disabled).

 

Ok if its horizontal/vertical, looks like at offset 03:E0: "90" is 1070 (230W), "91" is 1080 (330W). Someone needs to check if "92" is for 1070 SLI (460W). The GT73VR doesn't have 1080 SLI.
Someone with a 1070 SLI needs to post their EC screenshot (shouldn't matter if it's 6RE SLI or 7RE SLI, assuming the values are in the same location on skylake vs kabylake)

Yeah the 91 definitely worked. Can't be 100% sure if other values need to be changed also though. I did get a BSOD after like Firestrike Extreme finished (i'm guessing it finished, although maybe it didn't, it seemed like it was 10 loops but idk) might be because its too hot in my apartment, but when i rebooted, the battery hadn't drained at all.

 

Are you using RWE to change it?

 

Yeah

 

Awesome news! Congratulation of successfully beaten the evil battery draining and CPU throttling king! I guess soon you can release the full power of your 1070 while still having high CPU clock. The long fight finally paid off. Cheer!

 

I'm still not 100% sure if it's just that one value. There were a BUNCH of values different on my system vs Phoenix's system and I pretty much changed everything that wasn't read only. I think 04:E0 displays whether the PL is exceeded or not. My 20 loop test occured after I changed 03:E0 as well as everything else that wasnt locked down. But I was in a 70 degree air conditioned college library. HOWEVER, none of the values stopped the 230W power throttle until I changed 03:E0. But remember I changed a bunch of other values that were not read only, as well.

So I get home, into a 95 degree apartment (thats 35C), change 90 to 91, run 4 thread prime (AVX/FMA3 disabled) and firestrike extreme stress test, and get a BSOD after 10 loops. it completed apparently.

I guess something overheated, huh?

*Edit*
Ok something's up.
I finished the 3dmark 20 loops completely again, then got a BSOD with "IRQL_NOT_LESS_OR_EQUAL" after getting back to windows and futuremark collecting data or something. What the hell? LOL....

It was rock stable until it completed...I had time to see throttlestop window also...there was no throttling. like 1 second after I saw windows desktop...BSOD.

This is going to require more testing.
I have no idea why it would BSOD after finishing 20 loops...it didn't do that at the library...
If I abort the test after a few runs, it doesn't BSOD...
I'm 100% sure that one single value changes the power ID from 230W to 330W, but there may be some other value needed. A BSOD happening after a stress test finishes is 10 kinds of weird (4 thread prime was still running. Firestrike Extreme 20 loops had just finished).

 

Possibly a buggy driver. Synaptics, WiFi and GeForce drivers can cause it. So can too low CPU voltage.

 

Yeah I'll have to track that down. 8 thread prime didn't error out at all ...
So I'm going to leave 03:E0 offset as a preliminary answer for the moment.

I Just ran 6 loops then aborted FS:E with 4 thread prime running and no BSOD. So yeah idk. hot apartment, cpu voltage going nuts, I call it #firstworldproblems. The window AC is only powerful enough to make the living room "livable". I'll have to wait until I get access to another powerful AC again. It's been 110 degrees (43C) every day out here in Riverside, while Texas is getting hammered by torrential rains from that Hurricane turned into tropical depression....

k I can't do any more testing right now. The 330W PSU is running a bit too hot for my liking. It's actually as hot as the 230W PSU was in the library....gonna power off until I can get some cooler temps.