***The Official MSI GT80 Titan Owner's Lounge***

Leawo hangs and I could not get it to play anything other than the ads in front of the movie.
should I remove and reinstall the Power DVD? It takes to the place where you have to purchase it. I could not get one to start no matter what, it appears that GT80 doesn not even ship with it, in which case why is the link there?

 

Hmmm, sounds like another bug with Windows 10, you might check with Leawo support and post a bug.

It's been a while since I used Leawo, but I recall just like all newly installed Blu-ray software, there are updates downloaded the first few times you run it and play movies.

It's got to download playback codes, and new extensions on the Blu-ray playback for new movies. It still happens occasionally with new movies as the Blu-ray protections are extended.

This also happens with Cyberlink software.

I'd let the Leawo software load again, and let it sit there to download updates, then try again.

I'd also try right clicking and jumping to the Root or Title page from within the preview playback.

Also, I found using the Leawo software for DVD playback wasn't as pleasant as just using MPC-HC for DVD playback.

But, it does sound likely that Windows 10 has hosed playback from both programs for now...

Again, you could chase the tail of this problem for a long time without any gains, or replace Windows 10 with Windows 8.1/7, the GT80 came with Windows 8 so the drivers are there for both 7 and 8.1.

You could also follow the tips in that thread, good luck with that, but even if you get it fixed now, Windows 10 will re-break it or something else later with another Windows 10 update.

 

Can someone with the 5950HQ GT80 run an XTU stress test and tell me their results? I'm getting as high as 83C and idling at 50-70C afterwards and after a reboot. Wondering if i need to check the cooling and/or repaste.

Also getting a lot of Current Limit Throttleing.

 

Did you undervolt? Are you at stock speed, and did you OC?

Those temps are fine, the whole laptop holds heat so I recommend turning on the Full Fan during heavy testing, before and after to pre-cool and remove the heat afterwards.

If you have lots of M.2 drives, they sit on either side of the CPU on the top side - the CPU is right underneath, so they can hold heat and contribute to keeping the CPU hot.

I also run mine on top of a Coolermaster Ergostand III with the elevation set pretty high to keep the whole laptop at an angle to improve convective cooling.

To run long batch jobs I set the Multiplier to 35x x4 on the CPU and undervolt -120mV which really helps reduce the thermal output and keep the heat down for hours of compute time. If I am running long jobs on GPU's I turn on full fan too.

 

Just crashed without a bluescreen during the firestrike stresstest. atm i'm not OCing at all.

 

Is this a new computer or is this a computer you've had a while that's started having problems?

 

I've had this one a while, but it failing the stress test and crashing is recent. Is there a chance this could be VRAM overheating?

 

You need to change power limit to 45W however i dont know is PL can be increased on that processor. On the 4700HQ I know CPU turbo multi can be increased on all cores past default but I think you cant increase power limits on locked processors. You can try setting current limit to 100 amps but if that doenst work, the only thing you can do is undervolt.

 

You are quoting CPU tests, and now you think it's VRAM? What stress tests are you failing?

If you are gaming does this happen?

Were you starting to crash in games / apps you usually run, and now you are benchmarking on CPU and GPU to see where the problem is?

It would help to have some context as to how you got to where you are now asking these questions, like why?

 

Sorry, was referring to it crashing during 3DMark Firestrike Stress Test.

What sparked this was smelling something i thought smelled like burning, though it now that ive checked smells no different than my GT83 which i've never opened up once. (just smells like hot air i guess?)

I've had crashes and stuff months ago in games but that was when i was messing with overclocking, atm i'm running stock everything aside vBIOS, but even then i never had crashes without overclocking even with the custom vBIOS.

 

If you have both a GT80 and a GT83, are you spending most of your time on the GT83?

Maybe you aren't spending enough time on the GT80 to notice the progression of degradation in the GPU / VRAM?

I had a bad laptop Crossfire GPU once that developed over months, slowly getting worse until it died even when running stock, then finally completely went out.

Have you tried disabling one of the GPU's to figure out which one is the problem GPU?

 

Just crashed and rebooted during the sky diver stress test. now i'm worried. SLI keeps disabling during the reboot as well.


Most of the time i'm using the GT83, the GT80 lately just sits around acting as a minecraft server.

And i haven't tried disabling one yet, but in the thread about the burning smell it was mentioned that the VRAM may be failing since it has paste instead of pads, but since i hadnt had issues that i had seen in games i forgot that happened and only just now getting to fixing it. Could crashes like im describing be caused by overheating VRAM?

 

Yeah, especially since you tried to gap the void with paste instead of a thermal pad, the paste over time will contract and make less thermal contact with the heatsink, so it is the first thing I would check.

K5 Pro is designed to be used like this, and hardens solid without shrinking so it remains thermally good over time for bridging gaps when you don't have thermal pads available.

So, I'd stop using the GT80, order some thermal pads + K5 Pro and fix it before the damage becomes permanent...and you need to replace the GPU's.

 

I'm def okay going without using the cards but wanted to get to the bottom of the issue or i wouldn't be sleeping to well. Already got a razer blade in nonfunctional condition while waiting for parts to arrive to redo a LM paste job and a bad wirless card on the GT80 and now i've noticed this.

I ordered some fujiploy 1mm 17w/mk thermal pads for it this morning after remembering i did that to the VRAM. And i have antec 7 thermal paste and some kryonaut on the way i can sue to repaste the GPUs themselves. Will those work you think or is it already too late if its overheating and then crashing? (not getting any artifacting though)

 

I wouldn't even power it on any more, even if the GPU's aren't active they are still drawing power and the VRAM is powered on heating up.

It's not going to get better continuing to overheat them doing stress testing either

Just leave it off and then fix it, and once the VRAM is getting cooled when powered on you will know if the damage is permanent or if the cooling fix fixes it.

If the VRAM has been overheated for many months, as it sounds like it has, there is a good chance it's baked.

If you still have warranty, clean up the VRAM with stock thermal pads, not fujipoly pads, and use normal thermal paste too. Make it look as stock as possible for sending in for RMA replacement of the GPU(s).

That's why I don't recommend re-pasting MSI laptops, the gains aren't noticeable from a performance standpoint.

The MSI goop lasts a long time too vs. enthusiast pastes that need re-pasting often.

If something like this happens your re-pasting can be blamed and MSI won't fix it free under warranty. It's just not worth it.

 

I ran it actively for about a month like that i think until i got the GT83, and then only occasionally using the GT80 for games, it was mostly used for web browsing. Never had any crashes in games or artifacting, but now when ive put it through 3DMark stress tests it's crashing *unless cooler boost is enabled* which is why im hoping that it might just be that its crashing to prevent further damage?

And do they recieve that much power even when they've been disabled via the MUX switch thing?

Either way for now it's off and am waiting for the pads to arrive. I don't believe the laptop is under warranty anymore anyways since it was bought used.

 

The 2 year warranty transfers - but you need the original sales receipt or a photo of it. If the original owner registered it, you can create a login and have them transfer ownership to you, along with the warranty.
https://register.msi.com/home/login

Unfortunately the 1st owner is the only one that can do the upgrade trade-in.

If they registered it early enough they also got 3 free additional months of warranty.

It's worth following up on that warranty / ownership transfer especially if one or both of the GPU's are flaky.

The iGPU isn't powered when the dGPU is enabled, but that's under Intel CPU control. I don't see the dGPU's in the Device manager, so they should be powered off...but I wouldn't chance it

 

I don't have the pads that were in it when i got it, in a bit of stupidity in not knowing what they were (the one GT80 teardown at the time didnt have them) and there was a lot of stuff in there that it was obvious that the prev owner put in, i didnt think to replace them when they got damaged and otherwise messed up. I've since then discovered thermal pads and know what they are now, but since i hadnt had any issues in games i forgot i did something that stupid until last night.

 

Hey, it's all a learning process, hopefully replacing the thermal pads will make it all better.

The thickness of the pads is important, be sure and check to see if there are any gaps between parts and the thermal pad, double up where needed.

 

Will do, and i guess in the meantime also gonna research potential replacement laptops for if the GT80 does sputter out xx;

Though it's so hard to find comparable laptops when the GT80 has a mechanical keyboard. The GX700 has power but no mechy keyboard. The GX800 is ungodly expensive. And i have no experience with MSI outside this GT80 i got used and the GT83 i got on a black friday sale from iBuypower.

I'd love it if a single 1080 single PSU GT83 existed like how some of the first run GT80s had single GPUs

 

The new MSI GT75VR has a mechanical keyboard, and has a single PSU 1080 model.

The GX800 is available from Costco in the US now
https://m.costco.com/ASUS-ROG-Extre...ra-HD---G-Sync-Display.product.100339714.html

If you killed only 1 of your GPU's, and you can figure out which one, you can leave in the other one and get your wish.

 

The GT75 looks tempting if not for the lack of pricing info, but so far as far as ive ehard none of these chichlet mechanical keyboard hold a candle to cherry MX ones.

And i can't see the pricing on the GX800 there, but elsewhere it's like $6k, tad bit much to replace the GT80 ><

 

"Only" $5799, +55 for Costco membership.

 

Yeah... that's a tad bit out of budget. I may also consider just replacing it with another GT80 since they've come down in price, but performancewise something with a 1070 could probably outdo it without SLI issues and while smaller and cheaper.. but that keyboard qq

 

I got the GT80 SLI recently, here is my review - very nice.

the mechanical keyboard just awesome. the best one I ever had. And I use the same MX brown keys in my other keyboard.

the non-removeable battery is the dumbest idea of the decade. Especially only being 8-cell. A monster of that size needs at least 10, maybe even 12 cells. Given it is a desktop replacement. If you use a mobile laptop and you run out battery and no way to recharge... that's not good. It's nice to have a feature where the battery is instantly swapped a la Dell XPS.

the 1080HD screen is a bit worse than the same resolution on a 17" screen, of course.

All in all a very fast machine. SSD is a nice touch. I will end up reconfiguring it to use 4x2TB SSD.

I kept my last laptop until it was 6 years old. I can see keeping the GT80 SLI for a while. Even though technically, it's 2015 technology.


Eventually I plan to pick up a 4K monitor to run with it.

 

Please outline the steps to convert the laptop into 4x2TB RAID 1 configuration. i.e. 2 drives being mirrored by 2 drives.

Perhaps even like this. The boot drive mirrored with the data drive not mirrored but backed up every day.

 

Congrats on the new GT80

Not sure which model you got, if it's the original GT80 the 4x M.2 sockets are SATA - and AFAIK there aren't any 2TB M.2 SATA SSD's released - only PCIE x4 which are outrageously priced.

If you find a 2TB M.2 SATA SSD, please post the info here!

Also, if you have a GT80S MSI reduced the M.2 sockets to 2x M.2 PCIe x4 + 1 M.2 SATA, and dropped the 4th M.2 socket.

Please come back and let us know which M.2 drives you end up getting and how they perform individually and in RAID0

 

So i got the new thermal pads in but it still failed the firestrike stresstest with 89%. I noticed GPU2 is getting to 91C, is that normal? and is there a way to tell which GPU is which physically?

Rerunning with cooler boost on brought it to 99.6 passed

 

No, that's way too hot, depending on your room temp of course.

GPU should be in the 70c-80c range, CPU 85-90c range.

Did you re-paste the GPU's when applying the thermal pads? You can't just re-affix them without at least spreading the bubbles out of the paste - even it out at least - best to completely clean it and re-apply paste.

It sounds like maybe the pads are too thick and keeping the GPU plate from completely connecting with the GPU?

 

Afterburner isnt reporting a GPU2 temp limit which is weird.

I can try repasting again once i get some more 99% alcohol ordered. (EDIT there might be -just- enough left) I did forget to repaste the graphics chip thats on the right when the laptop is being used.

GPU 1 was around 78-82C, GPU2 was 91C

The pads i ordered were 1mm thick.

Repasted and GPU1 was around 75C and GPU2 was still 92C

guess ill try repasting the other one next?.

Wonder what went wrong bc i used the same pea method on both, and they both have the same thermal pads and whatnot. Any ideas?

 

You'll need to look for the differences yourself, look for fitment - "daylight" showing through GPU / VRM / VRAM to plate meeting, shine a bright light light and try to slide a thin card / blade between them.

Look for uneven tightening of screws. You need to turn 1 turn or a fraction of a turn on each screw all around - don't screw one down all the way, and move to the next doing the same, this tends to get one side overtightened and the others unable to compress enough for the plate / GPU to meet evenly.

I prefer to spread the paste thinly across the part and evenly across the heat plate as well - very thin, complete coverage.

I do scrape a perimeter around the edge of the CPU/GPU clean so expansion doesn't spill out too much - be sure and wipe any extra off as it will mess things up even if it's non-conductive it can look messy.

The pea method works to a point, probably better on desktop mounts as they have higher pressure.

There's a knack to it you learn over time...

 

>Look for uneven tightening of screws. You need to turn 1 turn or a fraction of a turn on each screw all around - don't screw one down all the way, and move to the next doing the same, this tends to get one side overtightened and the others unable to compress enough for the plate / GPU to meet evenly.

I wanna bet this might be it, ill take off the other GPU and repaste it again and be a bit more careful with the tension arm screws and hope it helps.

The pea method worked on the other GPU though it seemed so i doubt its my paste application itself..

 

If you think it's the screw tightening, you might just try backing out all the screws and putting them back in with small even tightening increments until done, and then check the temps. I think you can do that by just taking off the back again - you probably leave the case screws out anyway, so it should just pop off.

That way you can avoid one more re-paste, and verify if it's the screw tightening or not.

If it gets better completely you are done, if it's better but not the same as the other GPU don't sweat it - there is up to a 10c difference depending on the game / app - be aware of this. You shouldn't hit over 90c on either one though...

 

Yeah thats my concern, it's hitting 91C unless cooler boost is enabled, not even the 1080SLI GT83 with a 200Mhz OC on the GPU cores gets that hot. And last I checked the 1080s TDP was closer to the "990M" than the 980M?

Tried just the screws and no real difference, gonna try another repaste now.

Repasted and GPU2 is still at 92. GPU1 was around 74. I don't know what to do now.

I did notice a rainbowy sheen on the heatsink of the left GPU heatsink, could that be to do with it?

 

What are idle temps, on auto and Max Fan?

It almost sounds like there is a heatpipe or other heat transfer failure.

Or that GPU heatplate is still canted to one side...

When Max Fan is on is the air pressure output on both sides the same? Are both fans putting out the same volume of hot air?

That GPU fan should be putting out a lot if heat, if it isn't then the heatpipe or heat exchanger are troubled.

 

air pressure feels about the same.

GPU2 under stresstest with cooler boost on gets around 85C and the stresstest will pass at 99+ stability

Temps start on cold boot around 33C and after cooling down from the test GPU1 is 38 and GPU2 is 45


This would be a lot easier to diagnose also if i knew which GPU was GPU2 physically

 

Hmm, that idle temp looks fine, one is always hotter than the other, but it's unusual to have one get over 90c...

Do you OC or overvolt the GPU?

You may need to send this in to MSI to get it fixed.

 

Custom vBIOS but its never gotten this hot without heavy OCing, and both GPUs have the same vBIOS

There's currently no OC or under/overvolt

The laptop again doesn't have warranty anyways so i need a home fix if one exists..

 

Same vBIOS on both GPU's?

Voltage might be pumped too high as default.

MSI will fix laptops out of warranty, especially one as recent as the GT80, you just need to pay for the fix - but sometimes they fix it free anyway - especially if it's a hardware failure like I'm beginning to suspect.

As a sanity check you might flash back to stock VBIOS/BIOS.

 

Ill reload nV inspector and take a look at the GPU voltages


EDIT: they have the same voltage offsets

 

IDK, if you were reduced temps before, with stock vbios/bios, I'd try that.

Switch off SLI, enable GPU2, run a benchmark and see which side is hottest

 

Ok so GPU2 is the one i repasted second and that had the rainbow sheen on the heatsink copper.

ill try repasting again and checking the fit of the thermal pads i guess?.

 

You are approaching the point where you will need to rely on the realm of witchcraft and magic...

I can't think of anything else to try, except to start swapping out cooling parts till you find the culprit.

Or, flash the original vbios and see if that's it...

 

While i'm running the stress test, something i noticed was the heatsink copper is both has a soapy sheen on it and is scratched to hell. While i seem to be getting good coverage with the pea method im wondering if i need to replace the heatsink entirely.

 

You could sand it...but it shouldn't make that big of a difference. At one point it was working, right? And you didn't coat the heatsink. Conductivity over that large GPU surface should be enough.

Something isn't pulling the heat away from the heatplate, start checking for damage to the heatpipes, cracks, gaps, loose connection to the heat exchanger...

 

Its still getting to 91C.. its just taking longer to do so now. Firestrike stress test passed but it was still 91C by the end

 

That's with autofan, right? Does it seem to have enough air coming out? Do you have MSI DGC installed? Maybe adjust the fan curve for a little more spin? It's a bit noisy to run Full Fan all the time.

Or, if not, try MSI standalone Fan Control Software:

(Updated:2016/05/06)"Silent Option" Fan Control Application for MSI Laptops
https://forum-en.msi.com/index.php?topic=255972.0

 

Thing is if cooler boost is only getting it down to 85C anything less probably isnt gonna help. And yeah, thats with the built in fan curve.

Any other ideas what might be wrong? The main standout difference to me between the two GPU heatsinks is the overheating one's copper plate has a soapy coating of some sort and is scratched up quite badly.

 

Well, there are 6c between 85c and 91c, and 89c would be "livable"... if you aren't going to RMA to MSI, and need to find a way to live with it, a custom fan curve is a thought...

 

Getting it to 89C would still likely need max fans to push it and hold it below that, the fans ramp up quite a bit even keeping it at 91. Would like a second opinion on if there's any home fixes worth trying as i'd rather not be without the laptop for weeks and out hundreds of dollars if something like a heatsink swap would probably get it.

Removed another potentially problematic thermal pad and gave the GPU die a bit extra thermal paste, so far it's looking like one or both of those helped reduce the temperature difference a bit? Instead of a 10C delta from almost the get go it's only at like 5C difference now between the cards at two minutes in.