*** MSI 16L13 (Eurocom Tornado F5)/EVOC 16L-G-1080 15.6" Owner's Lounge ***

No because MSI doesn't have a machine ID programmed into its barebones firmware. It acts the same as if you wrote an "invalid" product ID into the GT73VR or GT75VR (like writing a value of FF or 00 into register "E3" in EC RAM (RWEverything), except that would cap the 7820HK at 45W at all times, rather than 91W.

It's a byproduct of MSI's cancer code.
Prema had unlocked it but he found out that the 16F3 would simply power off at high overclocks when playing 3d games because the VRMs couldn't handle the power requirements of CPU and GPU, so he left it at 91W so the GPU could still get full power.

 

Ah right. But thats a pretty good sweet spot too.

 

The curiosity was killing me, so I ran Firestrike. CPU at 4.4 GHz on all cores, GPU at stock for a score of 17,304: 21,374 in graphics, 14,054 in physics, and 8,316 combined.

 

those are good numbers.

 

Is it possible to run Linux on this thing? Say Ubuntu 16.04? Does nVidia CUDA and cudnn work well with the 1080?

Sent from my Pixel 2 XL using Tapatalk

 

there was someone with a tutorial on installing Linux on the F5 in this thread I believe

 

I think @syscrusher might have the answers you are looking for!

 

Install Ubuntu 16.04.3 LTS as usual and then turn off secure boot just for Ubuntu using

Code:

mokutil --disable-validation

Use this ppa in the link to install nvidia drivers https://launchpad.net/~graphics-drivers/+archive/ubuntu/ppa
Afterwards apply all updates and use synaptic manager or nvidia's .run file or deb file for CUDA 8/9 since Ubuntu 16 only has Cuda 7.5 in their repos. Use similar steps as mentioned in this link http://www.pradeepadiga.me/blog/2017/03/22/installing-cuda-toolkit-8-0-on-ubuntu-16-04/
I recommend you use linux kernel 4.13 update using Synaptic Package manager and don't worry about messages like DmC firmware or similar firmware missing error as you see in that previous link of cuda 8 installation. Intel firmware download section
Cuda toolkit includes most libs like cuBLAS, FFT etc..
Cuda 9 has more features like co-operative kernels and ML if you're interested. For best results use CUDA9.

 

Wow thanks! Is it easier to install CUDA 9 with 14.04? (Without the run file)

Sent from my Pixel 2 XL using Tapatalk

 

Use 16.04 and its lot better than 14.04. Since 17.10 is released most newer SW will target 16.04 LTS in coming years until 18.04 is released. Newer OS support Speedshift, better performance and supports NVMe out of the box.
Note: Bumblebee is obsolete and use nvidia-prime to switch between Intel and nvidia by logging out. Use Nvidia X server settings app which is installed during nvidia drivers.
On 14.04 default repos contain CUDA 6 which is quite older.
You can deb or run file as the latter contains all libs and packages for nvc toolkit for ML,Deep Learning etc..

 

So I am the same way. I am recreating this environment with the desktop. I have a Roccat lap desk with built in KB, and an extending monitor arm that I can swing over the couch/armchair where I sit. Then when not in use the monitor will swing out of the way and lower to be out of sight. Who knows I may decide to be back to laptops in 2 months for all I know. But this is a fun project for me and might let me pocket a few dollars from selling my laptop.

This all spawned because I got an amazing deal on a Ryzen desktop(yes a downgrade from the laptop power) for under 600 with a VR headset. After selling the headset I am under 400 invested in the desktop.

 

good on you for remembering who. Totally forgot who it was.

 

My wife would never go for me putting a monitor arm on our furniture haha. I like that the entire device is contained in one piece in the notebook. I bring it over to the dining room table once in a while.

 

I had a very similar couchtop computing setup. A monitor arm next to the couch and a laptray.

I eventually gave this up for a gaming laptop. The problem was that moving the arm and tower around was a bear and I'm doing a fair amount of traveling.

I still think there is a better couchtop solution that uses desktop component and is not a DTR laptop. I'd like to have more choices for monitors.

Sent from my Pixel 2 XL using Tapatalk

 

My wife is pretty cool. As long as I remove the arm when we have paries she is good.

Pixel 2 XL , nice. I'm rocking pixel 2 standard

But I agree I will insanely miss the laptop but I just want to do something different and I take my surface when I travel. And when I do travel it's for work and I have 0 time for gaming. I was hoping that would be different when I got this laptop but no.

Also I mainly play league of legends and my surface can play that haha.

 

Yeah it was a moment of weakness when I decided to buy this thing. Too much to spend on a phone...

It's quite nice to travel light. My old Ultra book was good enough to play old games on and I really liked it on planes. It is nice to have the monster books for longer trips and vacations.

Sent from my Pixel 2 XL using Tapatalk

 

After some more troubleshooting, I don't think the keyboard lighting software had anything to do with the power errors I was experiencing. As such, I've uploaded a copy of the software that was on my driver CD. You can access it here.

 

What is the SCM Drivers for?

 

So far I've only worked on servers with Tesla's via console. How does it work out using the GPU for display and compute? I suppose if I could use the on chip display driver while using the GPU for compute I would.

Sent from my Pixel 2 XL using Tapatalk

 

TCC isn't available on consumer cards and you really need quadro series for headless mode to work like in your office's server PC with Tesla GPU(maxwell/Pascal).
I was in your shoes last year when I multiple OpenCL platforms on Linux for my project where I had distribute workloads among CPUs and GPUs sadly Intel OpenCL drivers always overwrites nvidia OCL driver ruining nv cuda toolkit every time.
If you want best performance from ur GPU use Windows 7/10 using VS 2015. On Linux nv 1080 would be 10% slower.
I only tested it on 980M so your experience might be similar on Pascal.

 

Why the 10% hit on Linux? Windows isn't really an option - I need a real OS (j/k! The real problem is that the port to Windows is not feasible)

What I really want to know is what happens to the display when I'm running a big CUDA job on the GPU. This shouldn't really be a problem...

Sent from my Pixel 2 XL using Tapatalk

 

Nvidia driver performs best on Windows.
The display and compute will be handled on GTX 1080 itself so in case the compute gets hanged just restart x server using this trick https://askubuntu.com/questions/367983/how-do-i-enable-ctrl-alt-backspace-to-kill-the-x-server

 

so thanks to @Papusan that spotted a good offer ( here), I bought a Bitspower's ihs (skylake model) and then a 7700K for 310$ + 30$ for intel tuning protection plan, now yesterday after a PC shop delidded my 7700K I mounted all together and slapped with Conductonaut between die/ihs and on cpu/heatsink and today benched it a bit!

Removed "old"6700K


IHS to cpu die Conductonauted!!


and finally the Bitspower's IHS, Skylake version onto Kabylake cpu(seems fatter than the standard intel IHS)


7700K Cinebench@4,7Ghz 1.20vcore load(spikes up to 1,215v) 1049cb pts.


Aida64 Mem/cache Bench



wprime32 and 1024

 

Arrrg Cpu fan will not work unless the fans are at 100%. Man my new fans need to get here. I can't believe that basically every one of us had our fans fail at 1 year..

 

Cool that you managed to upgrade to 7700K & delidded with Bitspower IHS, congrats! Just looking at your pics, do you reckon there's a bit too much conductonaut going on there - considering there's conductonaut on both internal side of IHS & on the silicon chip itself, I'm wondering if there's too much applied? I know a lot of people like to apply to both internal surface of IHS & silicon chip, but I was under the impression that when applying to both you're supposed to paint on an ultra thin layer onto the internal surface of the IHS, but it looks like there's quite a pool sitting in the internal surface of the Bitspower IHS? Maybe it's just the pictures deceiving my eye. You might get even lower temperatures if you use a thinner layer of conductonaut.

 

yea shame that it happend, the msi forums are full of these complaints. i really hope the new fans can give me atleast 2 years.

 

Maybe is too much but I hope is only an optical effect, because I put a really small dot of Conductonaut prior to expande it.
Going to open next week because I finished all thermal paste for gpu(damn for this unified heatsink!) and I'll take pictures!
meantime..5.0Ghz Cinebench15 multi run

 

This is not optical effect.... All too much, bruh. Your IHS almost drowning in Liquid metal.

 

I too thought it was optical effect with light source shining on the metal IHS and die.

 

@Papusan Pappy dear, the liquid metal wasn't too much! (maybe some micron..)

Because you in the forum are my guides, I was really afraid (of spilled conductonaut) and finished to open again the IHS and heatsink (my poor gpu now has a tiny veil of thermal paste on it), I was cooking and my girlfriend bark at me that wasn't right time to do anything like this...I told her she was a pain and a matter of life or death! My motherboard or her, so we finished dinner very late to my fault.

Look at pics!

 

Some micron? Buy a new tool Or glasses JK
From my own application of Liquid metal.

 

That's WAY too much LM !

 

Hey man, don't get too obsessed with your laptop & let it interfere with your family life, forcing your girlfriend & yourself to eat late into the night so you can prove something to people on forums - I know it's easy to get obsessed, I can get like that if I think something's not right with the hardware, but I agree with your girlfriend ultimately! Does look like too much LM, Papusan's photos show noticeably less - in my two experiences of applying liquid metal I found the hardest part was getting the smallest blob of liquid metal to seperate itself from the syringe.

 

Its not only to prove but to not fck my Motherboard! The dinner were 25minutes late, I'm really fast after maybe 10 times already opened this laptop and also fast to cook and clean,.
Taking this apart, when opened it yesterday I removed a bit of conductonaut and today after work I'll remove another bit.
Thanks all for effort!

Edit:
Look too much also in this post, but no one noticed it or mentioned
http://forum.notebookreview.com/thr...6-owners-lounge.797128/page-653#post-10544073

 

(If you look at the post immediately after that one you linked I comment that there's too much paste ("white goo") - so it's not a conspiracy against you!)

EDIT: took me about maybe 45mins to an hour to apply liquid metal, including cleaning with Arctic Cleaning solutions (there's 2 seperate solutions) multiple times prior to applying (not including taking stuff apart and putting back together). I try & be as meticulous as possible because you only re-apply paste once in a lifetime (I hope with liquid metal) or once every year or so. Maybe take some more time to do it.

I'm not sure that removing a bit of conductonaut & then slapping it together again is a good idea - my understanding is that the best results are achieved when applying only once - if you break the seal then I think air bubbles are more likely if you don't start afresh with clean surfaces & a fresh application of paste.

 

Thanks @Papusan and @Robbo99999 @Falkentyne @Vasudev @Stress Tech
I don't like dead end spot conversation because I'm hoping to be here to learn first.. and maybe share anything useful.

Yesterday I was in rush and I've only repasted the Conductonaut without removing it, probably causing bubble issues, but the real thing is that I hate the "q-tip method" of Conductonaut and I prefer a lot more the "paint method"from CLU.

So here my morning "good"result (I think)

I started by removing with isopropanol alcool 99% the conductonaut on bitspower's lid and on the Tornado F5 heatsink

low Connection Alert for many images!

Spoiler

here the liquid metal regrouped an ready to leave!


my table!


heatsink's liquid metal regrouped an ready to leave!



Try to see how many pressure is applied by Bitspower lid



CLU!

PUBG 25minutes on my machine!



Recorded with no game audio to hear laptop sound!
7700K @ 4,7 Ghz max temp 67°c
GTX 1080 Underclocked @ 0,800mv @ 1508Mhz max temp 62°c,capped at 60fps by rtss.
Ambient temperature is 19°c and the laptop sit on a Thermaltake massive23lx at 800rpm


Hope is all ok now and have a nice weekend to all!

 

@aaronne I didn't say anything. Well I still think that big blob is enough to cover the CPU and GPUs incl. the heatsink too.

 

If newly applied Liquid metal is looking good... Just paint the used old one a second time. LM ain't as standard thermal paste

Too much standard thermal paste ain't dangerous. And he used a bad batch with Grizzly Kryonaut.

All good, my friend We don't like seeing fryed decent hardware
See this one
MiSJAH's guide on how NOT to Liquid Metal repaste!!

 

Oh man, you're not gonna like me for my comments here, but I'll preface it by saying that I'm not a delidding liquid metal expert so put more weight to people like Papusan & Mr Fox on this topic who have done it countless times, although I have done it just the once. You put Kapton tape over a small section of the exposed metal pins on the PCB of the CPU - now this is normally a very good idea if you're using the stock IHS, because the stock IHS does not make physical contact with that metal pin area; however the Bitspower IHS makes contact with the immediate area surrounding those metal pins, so you putting a piece of Kapton tape over it will cause the IHS to not sit flat on the PCB - uneven - I can see this potentially leading the IHS to not sit even on top of the silicon CPU chip. I know kapton tape is very thin, so you might get away with this, but I don't think it's ideal. The Bitspower IHS has a section cut away to protect those exposed metal pins on the PCB, so I don't think you need to stick anything over those metal pins when using a Bitspower IHS anyway.
EDIT: did some googling on thickness of Kapton tape, it's 0.1mm thick, so twice the thickness of a standard sheet of paper, not sure on how significant that is - I suppose it could mean you could end up with some gaps between IHS & silicon chip of up to 0.1mm, which also happens to be about the thickness of a human hair, and I know that during repasting they say it's important to avoid all contaminants such as stray hairs or fingerprints as this can effect cooling efficiency. Ideally you want as much heatsink metal to silicon chip direct contact as possible, with the paste just filling in the gaps on the micron level (0.1mm is 100 microns, perhaps too many microns). Having said that many laptop heatinks are wildly concave or convex anyway, so that flies in the face of micron talk.

Also, still looks like there's too much liquid metal on the interior side of the IHS.

Do you need to clean off all the remnants of liquid metal from the heatsink (so it looks like copper again) before re-applying fresh liquid metal to get the best results? I'm not sure on this, my idea of perfection suggests this is the case, but I think I've heard from Mr. Fox that this is not necessary & doesn't degrade performance, just putting this comment out there, not sure on that one. I also hear it's very difficult to get rid of liquid metal from exposed copper unless you sand it, so this gives me nightmares if I ever need to redo the liquid metal on my desktop GPU! (but I'm thinking I won't need to repaste that in the lifetime that I own it - fingers crossed!).

The amount of CLU on top of the Bitspower IHS and on the copper heatsink - I think it looks ok, maybe a little too much, especially as it looks like there is no protection around the CPU socket, e.g. the foam dams that Mr. Fox was using on his Tornado - to protect CPU socket from potential liquid metal being squeezed out & dripping. This is why a fair few people both in the desktop & laptop world only put the liquid metal between the IHS and the CPU silicon chip rather than using on the interface between heatsink & IHS, liquid metal also has the most benefit when used between IHS & silicon chip.

 

never spilled onto my system with 2 travel per year (over 8hrs plane travel) and I hope to not start after yours posts

quote from myself:

I'm pretty old school here, the CLU never dropped from my vertical desktop (when I 'd one..) so from 2012 I always repasted my and other laptops with no problem or issues IF the heatsink pressure was enough and the surfaces are flattened. Is like the CLU move all together like mercury if not contaminated. Never spilled at moment but I travel only twice a year, plus I cook my cpu with some daily hours of gaming! at 85°c CLU start burning itself and oxidizing, on m18xr2 I change it once every 2 year of use (from 2012) because heatsink has air gap.


2014
Alienware 18 CPU overheating

3rd repaste of 2017 after m18x r2 1060 upgrade
Alienware M18X R2 upgrade (lazy and actual guide)

 

I had to clean off the old thermal LM when I applied new LM EVERY time.

If I didn't clean off the old LM from the heatsink but did clean off and apply new LM to the BGA slug, temps would be skyrocketing and >10C core differences, because some of the old LM was hardened on the copper heatsink and had to be wiped off hard. Don't need to clean the stain fully, just have the heatsink flat and flush with no bumps. I did that and I was back to 2-4C core differences (5C if I was using AVX, 6C for FMA3 super stressful small FFT stress testing).

If I cleaned off the heatsink fully but tried to reuse the LM on the CPU even if I re-spread it evenly temps were not good, better than above but unhappy core temp differences even with no FMA3 no AVX.
So I have to always do both. fully clean fully fresh repastes after completely cleaning everything. Then I get proper results.

Maybe this is different for LM on IHS-copper (EDIT) HS, but I have BGA so there.

 

Yeah, so some people say it's ok to re-use and re-spread the liquid metal if it looks ok consistency wise, and others say to apply fresh each time - I suppose the safe bet is applying fresh each time.

With BGA the heatsink to silicon chip interface is absolutely critical, just like with desktop CPUs the IHS to silicon chip contact is arguably the most important contact area to get right. So, I can see re-use of liquid metal on the heatsink to IHS interface as being 'not the end of the world', but I see anything related to direct silicon chip contact to be critical - it's less surface area to transfer the same amount of heat, so that particular interface has to be spot on to avoid glaring inefficiencies (the way I see it).

 

There is almost no chemical reaction from Liquid metal between LID and die(no copper). LM between cpu Die and copper HS is something completely different

Edit. @Falkentyne See your own post above And this below

 

So will this issue with "cheap" fans on these units ever be resolved or not likely?

 

Unless MSI turnaround and acknowledge it as a manufacturing defect & then say that they have sourced a different batch or model of fans then I guess the answer would be 'no' unless through happenstance they move onto a different batch or brand of fans. Anyone know if MSI have acknowledged manufacturing defect & commented on rectification?

 

That's an aluminium heatsink though in video, not copper? Same reaction is not seen with copper.

 

It was to show @Falkentyne what would happened if you use LM on aluminum Intel and AMD's lid is pure copper coated with nickel alloy. Same for Bitspower's lid.

 

Uh.....remove the word aluminum. >_> Read that sentence (it makes absolutely no sense).
I think I wrote aluminum to mean the ALUMINUM FINS (radiator). i don't know. Sometimes I write things without thinking.

 

Ha, yeah, that's what you meant, copper contact plate of heatsink combined with aluminium cooling fins (fins which wouldn't be anywhere near the liquid metal).

 

Update: Just got my new fans via DHL from HIDevolution.

Still yet to open the parcel & inspect the fans.