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

I just tested that. It stayed about the same, maybe 10 points higher at 4GHz. 14556. I'll try overclocking with Afterburner sometime and see how it does, but I doubt I will overclock it playing games anyway. When I've done it before at 4.1 GHz with 200/350 overclock on the card I get 15673.

 

The games that used to crash, mainly PUBG and Quake Champions, crash no more. So far it looks like vbios 8A fixed the problem. Or at least put a bandaid on it.

 

I have been really busy lately, so is there any reason to "upgrade" my BIOS yet.

 

Not if you aren't having any problems. The 8A bios seems to fix a problem some of the earlier cards have with the boost clock. It "fixes" it by limiting the boost clock. I think you have the same card as mine. You can try the Ice Storm benchmark in 3D Mark. If you make it through it a few times without crashing you should be fine. If not, you may have one of the problem cards.

 

I'll look up that benchmark up next time I pull the mini beast out, I did get a random shutdown the other day while starting BF1 but after a reboot I had no issues for several hours of gaming.

 

I have a 6RE, everything I installed in it is provided at the MSI website, except for the latest NVIDIA driver 382.33. We have the same results per GPU-Z.

I had that problem before and was bothered by it because even off game its still at max memory clock, no games running and all, so, i searched the net. Solution was to uninstall all nvidia drivers and install it. In my case, I used geek uninstaller in removing all nvidia drivers (make sure to restart if the computer prompts you to after uninstalling) then reinstalling but this time I reinstalled the new driver (382.33) I havent tried reinstalling the one provided at the MSI website.

Hope this helps

 

Make sure you turn Gsync and Vsync off then try again. Had the same problem, Ice storm was crashing and @Hugodra suggested to turn off both gsync and vsync while doing the benchmark, I didn't realize my vsync was turned on at the NVIDIA control panel, I turned it off then tried again, fortunately, it finished its mark.

 

I get crashes, but only on certain games. I'm really confused. BF1 and BF4 run perfectly fine maxed out. Ghost Recon Wildlands crashes, unexpectedly and at completely random times.

Arma 3 runs fine, Fallout 4 and War Thunder fine. Random crashes with Far Cry 4.

Some of the most graphically intense games run better than those that aren't. I really don't want to have to return my custom to HIDevolution.

 

I think that was part of the game crashing problem I had. I ran Quake Champions beta and PUBG on low settings for better FPS on multiplayer. Doom on Ultra ran fine. I think the low graphics allowed the boost clock to go too high and crashed the games at random times. Just a theory. Since I installed the 8A vbios that limits the boost to 1645 GHz, I haven't had a single problem.

 

Same here bro, bought this from them. Have you tried using afterburner? I too dont want to bring this back anymore, too much of a hassle. Try decreasing your core clock by -400mhz when playing those games that crash.

As far as i've read, crashes in less graphically demanding games are caused by the GPU boost 3.0(shifting clocks or mhz every second) which is seen in the original VBIOS. Crashes happen because of boosting to a high core clock ,again, because of the GPU boost. Fix is through flashing the 8A VBIOS by removing GPU boosting or you can try what ive been doing, which is decreasing core clock through afterburner, since I dont have much experience in flashing VBIOSes.

 

Wish I had read this thread yesterday! Was debating whether to go for the 6RE or 7RF, ended up with a 6RE produced in 2016=crashing 1070 card.

Does anyone know if warranty is affected by flashing vbios with that chinese non msi approved one (8A)?

 

I'm not sure, but there are a couple of threads about 8A on the MSI forum and one person said tech support directed them to that thread. One other person may have said tech sent them the 8A bios but I can't remember for sure.

 

MSI support sent me the .8a VBIOS here in the Netherlands. However, I did not install it since my laptop works well with the one from the website. I suspect that the VBIOS is actually not to blame, but it could be the system BIOS or sth not properly setting power/voltage limits. This is a conclusion I arrived to since I used to run BIOS .113 on which my gtx 1070 reached boost over 2000 MHz. Now I am on .114 and with THE SAME VBIOS my card never exceeds 1885.5 MHz due to power/voltage limit being reached. One more option could be that sometimes the VBIOS does not install correctly - I unchecked windows hardware driver installation before installing the VBIOS this time under advanced system settings.

 

Heliada is your laptop created in 2016? Did you use to have crashes before updating bios?

 

and also I cant find any bios that end with .114, name of the bios on MSI support is e17a1ims.10c

 

Yes I have a 2016 model gt72vr, I'm afraid you have a different model and my bios would be useless, was just suggesting it may have something to do with it. It was installed by msi during warranty repair. Not sure I had the crashing before since I had multiple other issues, which msi kindly solved for me and exchanged my ram too. I sent it to them before I actually got to properly test it in games. But I know before the card would boost over 2000 mhz and now only 1885.5 on identical vbios, which is a clear difference and seems healthier to me since it avoids unstable clocks. In afterburner I can see it being on power limit most of the time under load and sometimes also hitting voltage limit. No crashes though.

 

Final Fantasy benchmark, can you share a link ? i dont know that benchmark.

Your card never goes above 1885mhz how? you are using 8A vbios?

Bios 113 and 114? what do you mean?

I dont see any bios or vbios, .113 or .114, they all finish with two numbers/lettters:

- 8A, 44, 0D etc

Are you talking about EC Firmware?

Mine is 1785EMS1.107, most recent one.

Low stress games are the ones that crash.

Ghost Recon Wildlands must be an exception, because its always low-stress games or benchmarks.

Its not a theory, thats whats happening.

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

So my laptop is back with new 1070 GTX, i tested many games, not a single problem yet.

As i said, the new revision of 1070 GTX fix this problem.

The values on GPU Z remain the same, so theres no way to find out the number of the version/revision of the card.

The only way to know, is the build number, if the laptop is from 2017, then you are probably good!

I will give more feedback if its needed, but for now, everything is explained:

- Laptops with 1070 from 2016: probably crash problems
- Low stress games are the one that crash
- You can fix it with 8A or 82 vbios, or -150mhz, that limits your system
- The real fix, new 1070 card

Thats it.

P.S

Vbios 44 official is back lol - https://www.msi.com/Laptop/support/GT72VR-7RE-Dominator-Pro.html#down-firmware

Ive read on spanish section of msi, that some laptops, windows does not start with .44 vbios, thats why MSI deleted this vbios from the official website.

Now is back again... what the?!....

Cheers !!

 

What was the turnaround time for you to get the 1070 replaced with the new one, how long did it take in total?
People say that MSI installed some pointless vbios that limits the clock speed than solving the real problem, is still still happening for future RMAs, and did it happen to you?

I just got the G73VR a few days ago and I'm having the same issues with the restarting, but only with DayZ Standalone so far. Do you recommend I just return it altogether?

Is there a specific way to know if I have the latest GTX 1070 card?

Thanks

 

Some people are getting full blown crashes with weird artifacts, all I have noticed so far is freeze+ restart. Should I return this laptop while I can? Has anybody tried the chinese vbios yet? What do you guys think?

Also does anyone know how to check if the GPU we have is the first revision?
I was watching this thread and both guys claim to have 11B7 as the last four alphanumeric characters in the Device ID section in GPU-Z?

Also in GPU-Z the Boost is shown as 1645 MHz. I'm not sure if this is correct, this was a used laptop I purchased but isn't it suppose to be 1900 MHz?

 

I ran GPU-Z in the game that would auto restart my computer for me. I had it OC slowly to 50mhz before it did. what do you think?
This was the log:

2017-06-28 01:37:45 , 1847.5 , 2003.4 , 49.0 , 1475 , 36 , 16 , 0 , 2 , 16 , 0.9500 ,
2017-06-28 01:37:46 , 1847.5 , 2003.4 , 49.0 , 1474 , 26 , 13 , 0 , 2 , 16 , 0.9500 ,
2017-06-28 01:37:47 , 1847.5 , 2003.4 , 49.0 , 1476 , 39 , 18 , 0 , 2 , 16 , 0.9500 ,
2017-06-28 01:37:48 , 1847.5 , 2003.4 , 50.0 , 1473 , 51 , 23 , 0 , 2 , 16 , 0.9500 ,
2017-06-28 01:37:49 , 1911.0 , 2003.4 , 50.0 , 1473 , 32 , 15 , 0 , 2 , 16 , 1.0000 ,
2017-06-28 01:37:50 , 1911.0 , 2003.4 , 50.0 , 1473 , 33 , 16 , 0 , 2 , 16 , 1.0000 ,
2017-06-28 01:37:51 , 1911.0 , 2003.4 , 50.0 , 1473 , 26 , 12 , 0 , 2 , 16 , 1.0000 ,
2017-06-28 01:37:52 , 1911.0 , 2003.4 , 50.0 , 1473 , 27 , 13 , 0 , 1 , 16 , 1.0000 ,
2017-06-28 01:37:53 , 1898.0 , 2003.4 , 51.0 , 1473 , 39 , 18 , 0 , 2 , 1 , 0.9810 ,
2017-06-28 01:37:54 , 1771.5 , 2003.4 , 51.0 , 1473 , 68 , 27 , 0 , 3 , 1 , 0.9000 ,
2017-06-28 01:37:55 , 1911.0 , 2003.4 , 51.0 , 1474 , 42 , 19 , 0 , 2 , 16 , 1.0000 ,
2017-06-28 01:37:56 , 1898.0 , 2003.4 , 51.0 , 1474 , 40 , 18 , 0 , 2 , 16 , 0.9930 ,
2017-06-28 01:37:57 , 1936.0 , 2003.4 , 52.0 , 1476 , 44 , 21 , 0 , 2 , 1 , 1.0310 ,
2017-06-28 01:37:58 ,

 

so it crash once it goes above 1.0v?

 

Could be, I didn't know. Don't clock speed and voltage go hand in hand.
Anyone to prevent this?

 

Final fantasy benchmark is just an official benchmark for the game final fantasy. http://eu.finalfantasyxiv.com/benchmark/

My card never goes above 1885MHz, with VBIOS 86.04.31.00.0d.
The .114 (E1785IMS.114) is not a vbios, but BIOS for the gt72vr 6re (your will either way have different versions as its different model). And no, it is not the EC firmware.

Low stress games don't even boost my GPU anymore (old wow from 2009 or sth stays locked to 1442.5 MHz on high performance mode, dips to as low as 900 or even 600 MHz in adaptive, while still pulling 180 fps - CPU is the bottleneck in my case). Previously in the same game the card would boost over 1900/2000 MHz in this exactly same game, using the same VBIOS. Something tells me VBIOS was not the problem or I installed it incorrectly the first time.

So in theory, something in my laptop changed despite the GPU staying the same and the VBIOS being the same, now it just behaves correctly. I ran ice storm benchmark on loop for 2 hours with boost speeds up to 1885 MHz and had no issues at all. There was clearly a power/voltage management issue before.

EDIT: in case you are wondering, no I did not go changing any settings aside from shoving nvidia control panel into adaptive rather than optimal and individual games into high performance mode. The clocks of everything are stock, the laptop just manages itself well. And yeah I did try the turbo mode in the dragon center, successfully adding +200 MHz to GPU memory clock (I did leave core clock alone tho) and CPU at 3.8 GHz stable... it did get a bit toasty when rendering a video (up to 70 degrees celsius on CPU being pushed to the max - but that's fine to my standards). With stock clocks the CPU never went over 65 or so. Overall, I am a happy customer now and LOVE this beasty machine.

 

So if everything is fine and dandy for you can you summarize what you went through, how the RMA process was, and if you actually fixed the issue or just put a bandaid on it? It would be helpful to other people such as myself who currently have this issue,

 

If you read this whole thread you will know what both me and Hugodra went through, we posted throughout the whole thing. Why Hugodra's laptop works is due to new GPU and why mine - well that's a mystery I would like to know too! Neither laptop has bandaids on it and I am sure Hugodra would most likely let us know if any crashes occurred again. My laptop just works now, it handles itself well. I do not overclock it though as it is pointless on this card - it already comes with factory overclock. If you check nvidia sheets the official boost is only 1645Mz. Now see what speeds your MSI is already reaching even without manual overclock. This card is NOT promised to be stable beyond 1900 MHz or even more, especially in laptop where you get limited resources (230W power adapter ftw). If it crashed on stock settings, that's another story.
@Hugodra what speeds is your GPU reaching now? I played the sims 4 for couple hours just to test it as a non-demanding game that actually uses the GPU... it held boost speeds over 1800 almost all the time, sometimes dropped to 1700's for a couple seconds but went right back up. No crashes. I think my laptop is just not affected (anymore). Still would like an explanation as to why my GPU behaves differently despite identical vbios as before. The only thing I can think of is the vbios was either corrupted or the system bios was not handling power/voltage well. One of the 2. Or MSI sneakily changed my GPU and lied about it - but why would they lie and also install the .82 vbios then. Makes no sense and it's a big mystery to me. I am just happy I can allow myself to love this thing now since it does all I ever wanted it to do and spent 2200 euro for (was my dream since like 2012).

 

My voltage goes all the way to 1.06 and it does not crash. But my clocks don't go as high as yours. Was that with manual overclock? Don't overclock manually. Just once more for clarity:

1645MHz is official boost allowed by NVIDIA. NO, it should not be over 1900 MHz. Anything past 1645 is nvidia boost 3.0, aka factory overclock which takes into account power and voltage limits. More load on GPU = more power needed => lower clock (for example). If these power/voltage limits are not obeyed, your laptop will crash which is what is happening with many people. The card is already near its max from factory. The fact you cannot manually overclock further does NOT mean you have a bad card, it simply means you did not get lucky to have a great one. Plus it is still a laptop with limited power adapter. If you crash on stock settings I recommend return the laptop unless you want a big headache. If stock settings bring no problems, then don't return it cause it works as it should. The end.

Article talking about desktop gtx1070, aside from minor differences it works the same for us:
https://www.reddit.com/r/nvidia/comments/5g9jo9/gpu_boost_30_how_it_works/
Attached is a short log of my the sims 4 gameplay to prove up to 1.06/1.05 voltage should be ok. Your clocks are probably just too high, as far as I see in this run my max was 1873 MHz and that was just a small spike.

 

God does anyone think MSI is stupid for not publicly addressing this problem. For such an expensive laptop?!?!

My clocks crash above 1885.5 it seems, with a voltage of 1.0620. I'm going to try some things you mentioned to see if I can fix it. And BTW this is NOT OC'd.

Mine also goes to 1873.0 like yours and does not crash at what point.

Going to do some more testing.

Another weird thing is the NVIDIA control panel crashes now when I try to open it, after I did a driver installation.

 

You can try this... Go to start, search "control panel". In there search for "system info". Click on the only result it returns (system). On left side click "advanced system settings" and under tab "hardware" click on "device installation settings" -> choose NO. Reboot. Download DDU + VBIOS from the website. Proceed removing all nvidia graphics drivers with DDU in safe mode as per MSI instructions and reboot. Install VBIOS. Install latest NVIDIA drivers from MSI website (only install the necessary and choose "clean install" or how they word it). Let me know. (Do NOT skip the stuff under advanced system settings - that is the only thing I did differently installing the VBIOS than before.) You can always switch it on later. If you still do not get good results, try reinstalling system bios/ec firmware from the website. Else I cannot help and you need to return the laptop or RMA it.

 

mine seems to be related to audio driver maybe. Mine crashes and then no audio until I restart the computer.

 

I ran DDU and just reinstalled the drivers normally using Geforce experience. The control panel works now.

I did an experiment. I disabled SLI in the control panel. I disabled one GTX 1070 from Device Manager. In the one game where I always crashed around 1850MHz-1900MHz speed, there wasn't a single crash+restart. Only thing that happened was a program crash due to memory error which later stopped. When in the game I would constantly crash 5-10 minutes in, I was able to play longer with no crash. My clock speed reached a max of 1870, and no crashes between 1800-1870 at all.

Pretty interesting observation... I'm wondering if it could be a power supply issue when, not enough for both cards, though I'm not 100% sure if DayZ Standalone uses SLI...

On aside, how much time do you think an RMA would take to have the cards replaced with the latest revision (NOT the stupid bios update...)

 

Well, my laptop went to Poland and it took like 3 weeks. You will have to get in touch with MSI for this if it is even possible to replace your cards. Does it still crash in SLI mode? Do other games crash? Did you reinstall the vbios too or just the nvidia drivers? It could be a power issue... You have the dual power bricks with this model if I am correct? Does the light on both come on/does one of them get hotter maybe...? One is sufficient for one gtx 1070. If one of them were faulty this could explain crashes when both GPU's are active since then both the power adapters are needed.

 

On a side note, it is possible only one of your cards suffers from this issue which is why it disappears when only one is being used. Strange things happen.

 

Other games do not crash, I have only had DayZ standalone crash so far. I've ran BF3, MW3, Black Ops 2, and mean to test a few other games. I just tested BF4 for 5 minutes and it went to 1898 MHz at 1.06 V without a crash - this is with both cards enabled from the Device Manager, but SLI turned off. Going to test it again with SLI on.
Is there a way to monitor and log both GPU at the same time for GPU-Z?
Edit: Just run the app again.

 

Both lights come on and both power bricks are hot... don't seem to be defective at this point. I still have stock vBIOS, my main concern is flashing the wrong one, like the one that turns off boost.

 

Off topic but do you also see the backlight bleeding in from the bottom when an all black screen is displayed?

 

If the clock speed for one card is staying at 1442.5 thru the whole game is that a good way to tell that the game is not utilizing the second card?

 

I would say yeah, if one card is staying on 1442.5 it means it is not utilizing it (or does not need to utilize it). If you only get a crash in that one game it could also mean it is game related and the game does not handle SLI well. Download 3dmark demo and try to run ice storm benchmark (the normal version, not extreme) and see if that crashes. Hugodra and others reported crashing on that benchmark even during single run, while I was able to loop it for about 2 hours (that pretty much means my laptop is stable and not affected). You can also try to disable the other card in device manager to test the other one too for stability when outside of SLI. If both cards pass through ice storm benchmark several times separately or even together, your cards are probably not defective. But if one of your crashes means a reboot too, it's not a good sign.

And don't worry as long as you download the right vbios from MSI support website for your model, it should not break anything if you follow the guides they give. You are looking at the 86.04.56.00.3A from the official website for all gt73vr versions with 1070 (no matter if sli or not it seems). As far as I could make out of it the only difference is that you need to confirm it twice during installation for SLI system according to some guys on other forums. To see which vbios you are using now, run msi help desk (should be installed on your system unless you removed it) and click export info under system info. Maybe you are already on the latest one.

No matter what, DO NOT install vbios 86.04.5B.00.44. That would cripple your performance locking the cards to max 1442.5. You do not want that. If you decide to send the laptop for warranty, make sure to explicitly tell that you DO NOT want that vbios to be installed. If they can't make it run right with normal boost speeds, they can replace the offending GPU. Also, do not install the 86.04.5B.00.8A as that will limit your card to 1645MHz (known as Chinese vbios). It will probably make your laptop run well, but it is not how it's supposed to be or what you paid for.

 

I have a different model than you. If it is just a little bleed it's ok, if it is really bad and noticeable while gaming just return the laptop if you still can. I am attaching a pic of my MSI in the dark (but I have the TN panel, not IPS) and pic of my asus from 2013 with another TN panel. Keep in mind camera exaggerates it A LOT. I never noticed how crappy the asus screen was before... well, I spent 2200 euro on this beast. Both are displaying the same 100% black picture. Needless to say the asus (on the right) was not all that black LOL.

 

It's only noticeable on pure black and only comes from the bottom of the screen. Does the 2017 model have this issue?

VBIOS Version: 86.04.2b.00.14,8388608

Might return it, we'll see.

 

You can use MSI Afterburner + RTSS setup OSD to display GPU utilization for both GPU's, and that will tell you for sure whether either or both are being utilized

I'm glad you found some help debugging your 1070 SLI; from @Hugodra 's experience it looks like you can get new 2017 GPU's from MSI, so that's great news too.

The screen bleed is something that MSI can also address when you send it in for RMA.

If you do return it and get a GT75, I'd wait for the GT75 to get officially released and reviews published with details before buying, rather than suffer from new model teething issues again.

Good luck

 

The gt75 will come with the new gpu revisions for sure. Shouldn't have these issues I think. If my laptop didn't work so great now I would've returned it and waited for sure especially since they advertise the gt75 with mechanical keyboard and man, I'd love that. Even if the current one is already great.

 

On my msi gt72vr 1070 laptop I have seen crashes/reboots in age of wonders 3 and then last night shadow of mordor after a few mins of playing. But then it was fine after reboot.
For AOW3 i updated the graphics driver to version 382.05 and it seems to work all the time. So prior to this thread it was hard to pinpoint whether it was a hardware or software issue or combination.

I'm on vbios: 86.04.31.00.0D from aprilish. In gpuz i will see my gpu reach around 1875mhz at a 1.06voltage. under performance cap reason i frequently see prw for power. so i wonder if we are underpowered by the power brick. gt73vr models have a bigger power brick

I ran the Final Fantasy benchmark on the default settings in windowed mode and i did not see any issues.

When it does crash in the event viewer you will see:
The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly.

 

I will say this to you and others:

- Dont try to search for a solution for the problem, you will waste hours and days, not worth it

- Contact MSI, ask for a refund, ask for a new laptop, ask for a new 1070 card

- Theres no way to know what its going on, we have some facts indeed, but no way to know for sure whats the real problem

- One fact is for sure: the 1070 card you guys have, its not good.

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

@mokifly ,

As i said before, yes MSI put a new 1070 card on my laptop, yes everything is working fine (FOR NOW)

Still, im not 100% sure that everything is fine, i need to play more games, test more benchmarks.

P.S

I will be back later, for more intel/answer.

 

The thing is I paid so much less for this laptop than the price it is sold new. I paid $2500 CAD after tax which is around $1600 EURO. The standard price is $3500 CAD before tax. So if I can fix this issue and keep it its great most of my games are not even affected. Amazon might give me one or two hundred discount for the laptop if I tell them the going problem so the price is cheaper.

 

Unfortunately more of my games are crashing now, it doesn't look good.

 

Why don't they make a vbios that limits the boost to 1800?

 

What model will be nice to get 427 or 017? or maybe 003?

 

how frequently are they crashing?

Sent from my SM-G930V using Tapatalk

 

It could be there is something else going on, not just or only the 1070 GPU issue, it might be something completely different.

I wouldn't waste time with it, either RMA it or return it.

It was a returned Amazon purchase with a high discount, those often have problems associated with them or they wouldn't be so heavily discounted.

You rolled the dice and it came up a loser, get it fixed under warranty or return it and move on.

 

Amazon offered me 20% off to keep it and RMA it. MSI says there is a center right in my province, turnaround might be around 2 weeks and all I do is pay shipping. I might go this route now.