** 1070 laptop: GT73VR, GT62VR, GT72VR reboot/crash problem **

As a first time user of windows 10, and first time owner of a gaming laptop. I suggest you check out this thread ( http://forum.notebookreview.com/thr...owners-and-discussions-lounge.794968/page-690) just read until the most latest post. learned alot from them.

 

Checked the box. Mine is a K1610N.

I have no crashing at all, ever.

 

Thank you for your feedback, i agree with you 100%.

Thats exactly what i been saying, why should i wait for a new official vbios?

To my understanding, to me, there will be no official vbios from msi to fix this, and if someday they release it, it will probably end up doing the same as 8A vbios.

Today, i received great news, after i insisted with MSI, they are going to replace my gpu for the new revision.

Do you know any way, how to check the gpu revision version?

Yes there are 1.0 GPUs with no problem and 2016 build laptops with no problem, like you.

But its confirmed that new GPUs or 2017 build laptops are working normal with original vbios.

I understand, why you and others like chinese 8A vbios, you can manually overclock to the clock you want, with the original vbios it will boost and is much harder to maintain a stable clock.

Yes, ive already shared what happened to me in the first post of this page.

They are not acting good, seems they dont care about their clients, simple as that.

Thank you for your feedback.

--------------------------------------------------------------------------------------------------------------------------------------------

Great news guys !

MSI is going to replace my gpu for a new revision.

Is there any way to check the GPU revision? maybe on GPU-Z?

Mine looks something like this:



If i was using a 1070 GTX revision 1.2, what would change on gpu-z?

Thank you

 

MSI probably

you can remove HS and check gpu it will say 1.0 or 1.2 right on the card. you will need to clean and re-paste.

 

Thank you for the answer.

You said "MSI probably..." is your sentence complete?

Remove HS, whats that?

Ok, but is there a way to know the revision without opening the laptop?

 

HS = Heatsink

 

Another crash for me in the witcher 1 today (capped at 60 fps with v-sync), with vbios 3A and driver 382.33. I'm quite sure now my problem is the same than Hugodra.

 

You have the same problem in your laptop, you dont need to test it anymore.

 

How did you manage to get your gpu replaced?

 

With persistence,

I had to waste many hours on online foruns such as this one, official msi forum and Zwame (portuguese) forum.

Will all the intel i could gather about this issue, i was able to confront MSI and they could not deny the facts.

Waiting for MSI to warn me when the gpu arrives i will send the laptop to them, so they can replace the gpu.

Now i need to know ways how to check the gpu revision, to make sure it really is revision 1.2

If you know someone with 2017 build laptop with K1702N , K1703N and so on, ask them.

I know @Falkentyne has a GT73VR 2017 build but he hasnt answered me yet.

 

Every single day i see new people reporting this issue, and MSI does not do anything.

Is this really affecting a low percentage of laptops? i dont think so.

Posted 23 hours ago:

 

Great post.

So what are you going to do now?

 

Great news, Hugodra. I told you you should take it up with MSI and bother them and insist on your case. And again and again until they either refund you/replace your gpu or whatever is faulty or give you a different laptop . Anyway, sorry for not replying earlier, but I sent my laptop for warranty (we don't call it RMA here) mostly because of my own retarded mistake (I formatted my ssd and reinstalled win 10, which got rid of the amazing factory CPU overclock), problems with drivers (due to win10 reinstallation) and crashing in wow with pink stripes (not windows related). MSI was so nice to take the 2-sth day old laptop in and perform extensive testing on it. It will arrive tomorrow hopefully. I heard they reinstalled factory windows 10 image, updated BIOS and replaced the RAM. I will not reinstall anything myself before I make a full image of the SSD anymore, lesson learnt - obviously I cannot handle it the same as my laptop from 2013. You don't wanna know what hell that MSI sent me through after windows reinstallation (e.g. loud popping from speakers being a new windows power saving feature - needed to be disabled in registry). Anyway, congrats once more, let us know how it works when it comes back!

 

Crash in wow with pink stripes looks like its the same issue as everyone in this thread.

Whats your laptop model? and gpu?

Thank you for the reply, yes i will give feedback when my laptop comes back.

 

I already mentioned it before over couple posts... but you can't really track all the posts haha. GT72VR 6RE-266NL (gtx1070+i7-6820HK), K1610 serial, VBIOS 86.04.31.00.0d. I would actually not mind having the GPU run at lower speeds, as it will still be what was promised according to official NVIDIA specs I looked up before the purchase and run at lower temperatures, thus possibly lasting longer.

 

Sorry, i didnt remember.

Well, msi usually installs vbios .82 that limits the card at 1443mhz, that is not what is promised on the official Nvidia specs. ( this is what is promised by Nvidia 1443-1650mhz).

Also, you loose Gaming Center function, Nvidia Physics calculation on GPU and lower performance by 5/10fps.

If your card use gpu boost function, it will never reach high temperatures, because the card lowers the clock if the temperature gets high.

Basically you gain nothing by wanting lower speeds limit by vbios, and you are helping MSI selling defective gpus on expensive laptops.

I know the problem is not that important, but for me is just not acceptable, ive paid for a top of the line laptop and i want everything at 100%

I just want to use it like my previous msi laptop, let it do everything by itself, control the clocks, control the temperatures, etc

Right now, im not celebrating or anything, after the laptops comes back and i do some tests, then i will celebrate.

Fingers crossed that everything will be Ok this time!

And, i hope you get yours fixed and everything running like you want,

Keep giving feedback, thank you.

 

Yeah 1443 doesnt seem right, but I think the Chinese one allowed boost over 1600, which probably would not then have the same effect as the 82 bios..? As in it would work as intended just with lower boost speeds? I don't know, but I do know I will not be testing or updating vbios or anything if it comes back working as it should. First things first I plan to kill windows 10 driver updates before connecting to the internet and only update in the future when absolutely necessary after a proper backup. I am starting to hate updates after what happened with my attempt at windows reinstall. Done the same thing to my asus like 4 months ago when putting in SSD.... was pretty smooth process (its newest drivers available are like 2013, but they don't bring any issues lol).

 

only way for sure is to open her up.

 

install my 1070 version 1 whenever my HS gets here and hope for the best

 

SIGH. So my laptop just got back from MSI. Card locked to 1442.5 MHz. **** my life. Now I am thinking if I should just install the official VBIOS from MSI website and try my luck. Or try the Chinese one. Sigh.........

 

If you use the locked VBIOS that runs the card at a specific clockspeed, can you overclock via MSI Afterburner?

 

I have the option to, but it has no effect at all. Neither does turbo mode in dragon center. Whatever settings I choose, nothing changes.

 

Plus sometimes the card would not go idle and actually really stay locked at 1442.5MHz even in ECO mode. This can be temporarily remedied by disabling the card in device manager and enabling again. See why I feel ****ed over? No one bothered to let me know such a change was happening. Feelings to you Hugodra, only when it happens one actually feels messed with so bad. My chance is to flash the official one on and hope for no crashes since I did not even have the chance to test it out that well before....

 

I heard you can manually over clock it. But not 100% sure. even at the base clock its an upgrade from the 970m by at least 70%

 

Yes the chinese boosts to 1650mhz, you can install it if want, it is working good.

Well, ive asked folks with 2017 build laptops, for gpuz screenshot, and it seems the ID Device is "B7"

Still need to confirm this

You cant send it back?

I told you and others, but you wouldnt believe me.

Yes you can with msi afterburner.

Dragon Center does not work with 82 vbios, only 8A chinese with 1650mhz boost

See? thats how i feel.

MSI didnt tell anything they just "lets do this and hope he/she doesnt know what happened, we dont want to change the 800 euros gpu"

You have 3 choices:

- ask for money/new gpu 1.2
- install 8A vbios
- install msi afterburner and use it to downclock the gpu boost

Yes you can, with msi afterburner.

If you have crashes while using a browser, then its a different problem mate.

Contact MSI, or see if you can return it.

 

I believed you, just didn't know how horrible it felt to have this happen. Anyway, just spoke to MSI on the phone willing to know what the purpose of this VBIOS is. He will let me know and said I should be safe to reinstall the one from the website. Think I will do that and see what happens. Anyway, I will pursue having the GPU exchanged if I get any crashes after that.

And no, I cannot overclock the card on this .82 VBIOS for some reason. It's just completely locked, not reacting to MSI afterburner at all. At least now after several updates of windows and dragon center it goes to idle... Too bad because they did solve all my other problems with the laptop, just introduced another one/tried to cover one up.

 

when utilizing the VBIOS that locks performance clocks to 1400mhz range (or the chinese 1600 range), and being overclockable via MSI afterburner, does this vbios also crash at high mhz? Like if you "overclock" to 1900mhz, does it also crash?

 

this is my understanding of it. there are the unstable card that needs a Chinese vbios 84 or something that will limit it to 1600mhz. stabilizing it but not allowing it to boost past 1600mhz. That card if you don't flash to the Chinese vbios you can manually over clock it. not sure what the overclock #'s are. and not crash. the problem with this card is the boost clock. I guess manually overclocking puts it in a steady frequency which makes it stable because its not using the boost clock. Now MSI provided there vbios 82 or something that locks the card at 1600mhz and prevents you from manually overclocking. then there is the MSI version 1.2 card. I'm not sure but I think this card works fine no issues meaning the boost clock works the way its suppose to but not sure if the overclocking ability is locked or not. not enough info yet. hope this helps. This is my understanding of the facts and all may not be true. I'm just a dumb hillbilly.

 

The .82 from MSI warranty locks it to 1442.5MHz, which means NO boost at all and no overclock possible (at least for me). Chinese VBIOS does allow boost from the basic clock (1442.5) up to over 1600 MHz (not sure how much exactly). Overclocking is possible with the Chinese one unlike the MSI warranty one. Stock VBIOS from MSI website allows further boost all the way up to over 2000 MHz, which leads to kernel power or other failures for some users with the 2016 builds (gtx 1070, revision 1.0). These builds start with serial number K16.... Some cards are even unstable with clocks over 1600 MHz and should be replaced, it is clearly a faulty piece of junk not worth the money we all spent on it. I will pursue getting my GPU replaced just as @Hugodra did if it turns out my GPU is also faulty. I still don't want my money back cause I really like this laptop and I think the keyboard and sound are really great. Plus it runs decently cool...

 

I didnt know the 82 vbios cant be overclocked with msi afterburner, well i didnt try it, but makes since since shift mode on Gaming Center does not work too.

 

Alright, so are all K16 based serial numbers afected 100%? Or are those serial numbers the ones more common to be affected? I have a GT73VR with a GTX1070 and a serial number K1610N. It is still new in box, so I would rather not open it, but if it does end up being affected... sigh I will have to deal with it anyways.

 

GT73VR 7RE or 6RE?

No, just some of them, you could in fact have a 100% working laptop, theres no way to tell, you need to open it.

In my country you have at least 15 days to test a new product, if its defective you can ask for money or a new product.

In your country is not the same?

If you indeed open it, give feedback, take a screenshot of GPU-Z as well, test it with everything at default (bios,clock,etc)

Good luck

 

Thanks, if I do decide to open it, I will definitely give feed back. I will also try to help by sending feedback to MSI as well.

 

@Hugodra Hey there! Do you still have the laptop with you? I got official response from MSI on which VBIOS to use.. guess what? It's the chinese one, they actually sent me the file. So I went ahead and was like **** that, first I am trying the one from the website. I am just wondering... how long could you run ice storm benchmark before crashing? I have it on loop now for past hour, just made it full HD and bumped up the settings a bit (like MSAA to the max etc). I noticed though, that the 1070 now does not have the weird over 2000 MHz spike as before it pretty much maxes out at 1885.5 MHz and seems stable so far. I also ran firestrike before and gamed wow for an hour. Wonder if it's maybe the combination of VBIOS and system BIOS and nvidia drivers or sth that makes it run well? I am on BIOS .114 now, which is NOT the latest one (MSI installed it for me), VBIOS .0D and nvidia drivers from the website too (375.63). I do use GSYNC as I will be using that when playing anyway. My graphics card is set to optimum performance, thus not maximum. I don't know... but it works like a charm now? Maybe I was lucky to begin with and only crashed due to the faulty RAM they replaced. What do you think?

PS. It is not throttling due to temps or anything which would explain the lower boost speeds, since it stays below 70 degrees and the fans are actually a LOT quieter now too. I don't know anymore... pretty happy with it if it stays like this.

 

Hello, I have the GT73VR-7RF with GTX1080 (K1611N)... Am I affected by this problem? I do crash sometimes... could be overload. Also, please check this: http://forum.notebookreview.com/thr...d-from-ac-adapter.805591/page-2#post-10542039

Br,
Ralf

 

What kinds of crashes? While doing what? What do you see in event viewer when it happens? Does any monitoring software point to voltage/power limit being reached at that point or both of them? My card is reaching power limit pretty much all the time when pushing it to 100%... but so far no crashes after about 15 hours of gaming (turned off sound in wow and no freezes anymore either so it really seems it was only sound driver related). Also went through at least 2 hours of benchmarking (unigine heaven, firestrike and icestorm, all passed with flying colors).

PS: Does anyone hear coil whine from their laptop??? I can hear mine whine/chirp when the GPU is not being stressed.

 

@Hugodra Ok just a quick update... my laptop seems to work 100% now, apart from some coil whine which appears is normal and also happens in other brands like asus laptops. I monitored closely how the card behaves now, it will not boost if it cannot anymore - during final fantasy benchmark the core speed drops as low as 1500 MHz as it constantly dances around power limit and actually hits voltage limit many times (sometimes both at the same time). I ran the same benchmark on my boyfriend's PC with desktop MSI gtx 1060 and it behaves the same, though it keeps boost speeds better since it is heavily bottlenecked by his ancient 6 core CPU so it actually does not get utilized to 100% as often as mine does. My score in that benchmark is over 12000 and his over 7000 so that seems about right. My card also NEVER goes above 1885.5 MHz now. Either way, I am keeping the laptop. MSI support in my country was pretty helpful aside from advising me to install the chinese vbios. MSI should have a talk with the guys in Poland tbh and work their stuff out as installing the .82 vbios is a bad move, especially if there is no reason for it.

My card is now set in nvidia control panel to high performance for games and adaptive otherwise, GSYNC on. Works well it seems. Wow freezes with sound on thus incompatible with realtek, but laptop itself does not reboot or freeze. 0 issues with other games & benchmarks (dragon age inquisition, final fantasy, many benchmarks - icestorm, unigine heaven, firestrike,...)

Attached is a screenshot of the card behavior as noted by MSI afterburner and my system info. Hope it helps someone, at least to further prove some of these laptops can work without crashing.
SN K1610N...

 

I wonder if this could be my problem? I have the K1608N made in 8/2016. It was working fine until a couple of months ago and then when playing a game it would lock up, get artifacts, freezing audio and then reboot. Quake Champions - often, PUBG - often, Mass Effect: Andromeda - maybe once or twice but I can't remember. DOOM, I don't remember any problem. It does it overclocked and not overclocked. It might do it 3 or 4 times in a row and then not for several hours. It has been driving me nuts trying to figure out what is causing it. I tried the older nvidia driver from MSI, 376.39, and it doesn't reboot so I can see the event. It is always Display driver nvlddmkm stopped responding and has successfully recovered. Using the newest drivers it locks up with artifacts and reboots and doesn't write an event. Quake Champions and PUBG are basically beta games so I thought maybe that was causing it.



Also, why does HWiNFO show 2000MHz memory clock and Afterburner show 4006MHz?




Also also, does anyone have the original vbios for the 6RE? I want to test something.

 

What are the highest gpu clocks you get in games/benchmarks? Here you only show hwinfo for idle I suppose, that is not helpful. According to what Hugodra found, you may actually have the second revision of the graphics card (ends with B7 in gpu-z)... anyhow you should call up MSI as it seems it is likely a hw fault. If you do send it for repair, make sure to specifically tell them not to install the debug vbios (ends with .82) as they cannot test it with it. Else you can try to install this vbios: http://www.filedropper.com/86045b008a (shared to me specifically by MSI support for benelux here in the Netherlands) - that is if you can live with limit of 1645 MHz (not like it goes higher in games mostly anyway). - I did not upload this, MSI support did and shared via email so it should be trustworthy source.

 

Thanks, I already downloaded that vbios today but haven't installed it yet. I was looking for the original vbios that came with the laptop just to test and see if it made a difference. This is HWiFNO playing DOOM and it was the same when I tried Firestrike except the load was 99%. No problem with crashing with either of those.



It is surprising I might have the 2nd revision of the card. I bought the laptop within 2 weeks of it being available in August.

 

The card should be stable at these speeds. You should try some less graphics demanding game or ice storm benchmark to test if the clocks will then go over 1900 mhz or even more as that is not okay. Oh and memory clock in hwinfo is the same for me, nothing to worry about. Afterburner shows the correct memory speeds. If you crash in ice storm benchmark you have a problem. First try to set high performance in nvidia control panel under manage 3d settings, then see if you crash again.

 

I always run it with prefer high performance in nvidia. I tried ice storm and it locked up and rebooted in about a minute. So does that mean I have the card revision problem or what?

 

I think you could. Artifacts are never good sign. My laptop runs well through all benchmarks I throw at it ans that's the way it needs to work. On the other hand you have the driver stopped working error, not kernel power error. I'd say contact msi in your country, explain what's happening and explicitly say you do not wish to limit the cards clock by a vbios (.82 or .8a). Let them first offer other solutions. Try to convince them that if they cannot make the card perform as it should that you want it replaced for one that can or your money back. Insist until they agree, that's all I can say. Don't let them install a vbios that hides the problem.

 

I read on the MSI forum someone used Afterburner and set the core clock to -150 and that fixed their problem. I tried it and made it through Ice Storm with no problem. I'll do some more testing with it in games that crash.

 

What are the max GPU clocks with the underclock? It could be just a temporary solution as I remember reading from some people that problems returned later. The VBIOS 8A seems to be the better option, it can even hold overclock for some people above 1900 MHz.... if you feel adventurous enough to be underclocking I would recommend trying to install that first instead of your current VBIOS.

 

I'm going to try 8A tomorrow. I tried Ice Storm again at -150 and it locked up this time. If I was reading the Afterburner onscreen overlay correctly, the boost clock was reaching 1898MHz.

 

1898 is too much I think, mine never goes beyond 1885.5 and I have no issues. Still not sure why this is, already gave all info I could to MSI to figure out why some of these cards reach unstable clocks. Something that happened to my laptop during or after the warranty repair made it run well, not sure what (I had the same vbios as now before and it would actually spike over 2000 MHz on core speed, which is a lot even for full desktop cards).

Actually, my card now does not boost at all unless needed (old wow 3.3.5 from 2009 or sth runs on 1442.5 MHz (up to 20% load) and goes up to 190 fps (CPU being bottleneck due to bad optimization), while dragon age inquisition/final fantasy xiv - relatively new - boosts the card to its limits which is mostly around 1600 MHz under full load, or over 1800 when sitting at 40-50% load... To me that seems like a healthy behavior. I monitor all speeds etc with MSI afterburner overlay in games.

 

I am glad your issues are mostly attended. It makes me more hopeful that if any problems appear on my machine, at least there is a combo of VBIOS, BIOS etc that I can try to make it work.

 

I just finished several hours of changing testing changing testing changing testing.

I went back to the vbios my card shipped with, .14, from the currently offered MSI vbios of 3A. Made it through 1 round of Ice Storm, boost clock would reached the high 1800's .. crash .. reboot. Lowered CPU from 4 GHz to 3.7 GHz, applied Gsync .. 3 rounds .. boost clock would reached the high 1800's .. crash .. reboot. Flashed the 8A bios, back to 4 GHz CPU, no Gsync .. 5 rounds of Ice Storm .. boost clock never exceeded 1645 MHz .. no crash! My Ice Storm score was actually much higher with 8A, 164853 compared to 125419 with .14 and .3A vbios. I haven't tried any of the games yet that I've had crash/reboot but I think this will fix it. I'll update if it doesn't.

 

Great news! I guess that seems about right, I get 130000-ish score on ice storm with cpu at stock speed (3.2 ghz or something with all cores active). How's your firestrike score now?