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

It was never All GP104's that had the issue, only MSI's 1070 on their custom PCB design, which most likely means it was MSI mistake

MSI RMAed and replaced the cards from people who complained. Then sold the borked GPUs to second hand dealers who sold them to people like me.

Who knows how many owners are stuck on the 8A no-turbo vbios as a "fix". The issue is moot though as the real fix, to stop the cards going above 1.013V, is no limiter at stock specs, if anything a light bump in performance due to undervolt, only at extreme OC well over 2000MHz / 150W.

 

try using 86.04.31.00.OC vbios you can go with a 200+ core clock and 300+ memory clock just curve at around .950mv

 

I hope people demand MSI to fix the issue or money back / new laptop, make your rights count folks!

 

At least, you can check in hardware if you have the good GPU, in my case i have the version 1.2 that is the new GTX1070 without the boost problem.

 

Great information my man @Pedro69 !

Good job!

 

Interesting thing happened today, i underclocked the core clock and memory clock on my 8A bios flashed card and it allowed the voltage to go above the crash limit of 1.013volts image showing it hitting 1.05 volts any ideas why this happens. maybe we can find a fix.
https://drive.google.com/file/d/1pSf8HiVyyYJ5R5A807WJEXDaDpe1CevB/view?usp=sharing

 

I just got my GT73VR 6RE and of course I have a GPU from October 2016. I see that the previous owner has already flashed the BIOS as well. If I add a +200mhz OC, will that at least get me back to the default clock speeds this thing shipped with originally from a performance standpoint?

 

hey man, welcome to the club. in all honestly a 200mhz overclock should get you to 1845mhz. a normal notebook 1070 will boost to just under 1900 under good conditions. so you can expect similar frame rates to a regular notebook 1070. one positive about the locked voltage and overclock is stable frame rates. because the clock-speed isn't constantly shifting your fps should be very stable. also the lower voltage will produce less heat so your gpu fan should run at lower rpms and stay quieter. dont worry about it too much I know a bunch of people complain about it but i think the pros far outweigh the cons. I much prefer quiet operation and stable fps then 50 more mhz

 

Thanks man. I am coming from a fully decked out Tornado F5 so trying to gauge my expectations accordingly for what I paid for it. I'll give the OC a shot once I get everything setup, previous owner blew away the recovery partition and had no idea what they were doing (installed Windows on the mechanical HD and didn't even have the SSD formatted).

 

why did you change from a tornado f5 to the gt62vr?

 

I paid a lot for my tornado and had some medical bills for my wife pop up around Christmas. Sold it to help fund Christmas for the family and I have a fairly nice desktop so it was essentially overkill. I am very much looking forward to the F7 when/if it comes out.

Side note - max OC on my unit is +250 core and +115 mem without the video card going to a black screen. (1885/4119).

 

stress test that, thats right around the max core ive seen, however ive seen memory go much higher then that. might try dropping the core a bit and pushing the memory more.

 

I'll throw some more at it tomorrow. Didn't push memory just yet, want to make sure everything is full stable.

Side note - does the 8a bios remove gsync certification? I do not see it anywhere in the options list under the Nvidia control panel. Confirmed running at 120hz.

 

g sync works perfectly on my unit. if you cant find it in nvidia control panel, reflash 8a and do a full reboot. might need to check out display driver uninstaller as well.

 

So you actually have a gsync option under the display portion in the Nvidia control panel to turn it on and off? Hmm.

Thanks for all the help.

 

Can someone do me a favor and tell me what mobo you have with the 120hz panel? Did they make the 6re with a 4k panel? Wondering if it was swapped.

 

Like my nickname... After a month, laptop returned me and nothing changed or solved. Same gpu, same vbios and crashes again.

 

Some guys says Revision A1 and 2 graphic cards but i don't see any A2 cards anywhere. What difference btw. faulty and non-faulty cards? Launch date? Or transistor count, boost speed, vbios, device ID?

 

So happy I found this thread. Have the GT62VR for about a year now and having the same restart issues. I never realized it could be an issue with the laptop itself and thought it must have been like Windows or software problems. I am going to send the laptop back to MSI, but I am only allowed to give a description of 300 characters. To be honest I do not fully understand everything in this thread (I am not as technical as some of you guys here). Any advise on what best to but in the 300 characters to MSI to explain the issue? I understand from here that most likely they will update the VBIOS, which means that I will have slightly lower performance? To be honest, if it isn't a huge difference I might be ok with it, as I won't be able to fix anything like this myself, and I just want the crashes to stop .

Thanks a lot!

 

Not true (for me)

 

So... I've been using my GT72VR 6RE Dominator Pro Tobii for a while already

I had the constant-crashing issue... So i used the "Fix" 8A bios, now i'm reading that the "fix" is not good at all, so.. what do i do?

Is there any better bios? I've been seeing people talking about voltage... I cannot touch the voltage on my 1070 on MSI Afterburner, it is grayed out :S

Buyed my laptop on january 2017 so for sure it has the problem.. should i RMA?


How should i make the curve? can anyone post me a picture? never used it before

 

If the fixed 8A BIOS works for you than previous one you had then by all means keep it. If MSI cannot see the problem then they will update the firmware on the card and then return it back to you unless the card itself is truly defective.

 

I am in desperate need of help, i don't know where else to look and this place seemed to have the most information on this. I don't have the laptop models mentioned in this thread but i do have the original Vortex G65VR 6RE model where the defect originated from. I have already went through the tiresome trouble of RMA but it ended in nothing being done. I have tried changing the voltage with MSI Afterburner but it won't allow me to no matter what settings i change. It also doesn't seem that the "8A" bios is compatible with my pc, as it tells me it is the wrong file size. Is there something that can be done? I have also attached a GPUZ statistic.

 

Do you know how to extract files properly?
"Wrong file size" has nothing to do with the vbios and everything to do with not extracting files correctly.

Go download the free trial version of Winrar.
Then use that to extract zip and rar files.

Then use the newest NVflash to flash it.
here is the .8a vbios

And you can't CHANGE voltage in msi afterburner. As has already been discussed multiple times (did you do a search?), you have to LOCK the voltage point on the graph with the L key, and hit apply.
Control F brings up the voltage frequency curve.

 

Thank you for pointing me in the right direction, but i am still having some trouble. I have attached another image displaying the problem.

 

If your system does not have a gsync video card, you cannot flash a gsync Bios onto it. The 8A Vbios is gsync, as well as .0D and .44. Before, you could override the error with nvflash -4 -5 -6. I assume that doesn't help now?

 

Unfortunately no. I have even tried the -j -f --protectoff parameters.

 

@Nemesis90 does your card support gsync or not?

 

Yes, an GeForce GTX 1070. It was in my first post.

Edit: Actually i forgot that it is the mobile version, i'm not sure if this complicates things.

https://www.techpowerup.com/gpudb/2869/geforce-gtx-1070-mobile

 

You still did not answer my question.

 

No it doesn't look like it, it doesn't seem to be under 3D settings in the control panel.

 

1st question, what version drivers are you running? Have MSI updated signed g65 drivers beyond 375.70?


Do you have a gsync monitor?

You don't need to flash 8A vbios.

2 possible remedies (I have two ex-vortex 1070s in my P370EM project thing) both are done in afterburner, and if you save in a profile and set it to load on startup these are set-and-forget settings

1.
Flatten the boost curve from 1.013V through 1.062V in afterburner so the MHz is the same on all those points.

The card just doesn't need that extra voltage until pushed to desktop 1070 OC levels, >2000mhz/>150W. 1.013V max is fine for stock 115W you will perhaps see a slight uptick in performance due to higher efficiency

2. As you were told, lock to a specific point on the curve.

But the card will idle at those speeds eating 30-40W instead of ~6W that may cause fan noise

Its the transitions across the 1.013V threshold that cause the crashes, either method stops them. The 8A vbios does it by completely disabling turboboost, the nuclear option.

 

Ok, you need to find a non gsync vbios. However the 8a vbios is gsync only, and it limits the maximum GPU voltage to 0.881v.
The "buggy" cards crash at the 1.013v voltage step. It's unknown why but it is known that something is wrong with 1.013v onboard the card itself.

There is no "non gsync" vbios that limits maximum voltage to 0.881v. You can use MSI afterburner and manually lock the 0.881v voltage point yourself. I told you how in the previous post. Then you will never encounter that crash. You can save your locked voltage profile to an Afterburner preset also.

Or do what @bennyg said and flatten the curve. That works also.

 

Damn, I'd be pissed if I spent a bunch of money on a nice laptop in 2016 and it still has major ongoing hardware issues.

 

Or $2.5k - $4k on a small form factor pc.

 

No i do not, this is running from a TV display. My driver version is 391.35.

Is this what i am supposed to be doing? This is from 1.012V through 1.062V, there is no preset for 1.013V.

 

The voltages can sometimes vary slightly. But start it before the 1.012v point.

Flatten everything past it

 

Yes that's it. Don't need to go beyond 1.062V. Keep that window open while you run something and the crosshair will show which point the GPU is currently on. It shouldn't go past the lowest voltage point on the flat line. It's worth double checking since the UI is horrible and you could miss a point that's +1 higher despite it looking perfectly flat.

If your display isn't gsync capable the option wont appear in nvidia control panel. But might still be a gsync capable card. Run GPUz or something that shows the device ID. 10DE-1BA1 = non gsync. 10DE-1BE1 = gsync.

Where did your 391 drivers come from? If MSI website can you please give me a link?

 

He posted the GPU-Z already.
it's non gsync (1BA1)
https://ibb.co/iyATwn

 

GPUz shows the ID starting with 10DE-1BA1. I got the driver from Nvidia's website, the latest driver from MSI support is 391.01.

http://www.nvidia.com/download/driverResults.aspx/132841/en-us

 

Well everything seems to be okay now. I set the frequency to 1695Mhz (0.881V) since flattening the curve from 1.013V through 1.062 didn't seem to work.

Thank you two for helping me and being patient with me. I will let you know if there are any other problems in the future.

 

hope this doesn't happen to me I'm looking to buy a new msi machine in 2 days. Which will have a gtx 1070

 

Hey man, having some random reboots with my GT72VR and it looks like I'm in the same situation as you - can you show me what you made your graph look like for the voltage in Afterburner? Or maybe upload a preset you saved to get yours stable? The constant reboots are killing me. I'm messing around with Afterburner tonight, so hopefully I can figure it out on my own, but I'll monitor this page and see if you or anyone else might be able to help me out. Thanks!

 

Flat curve: see post 386 in this thread http://forum.notebookreview.com/index.php?posts/10711162

Lock: select a point and hit Ctrl+L. Clocks and voltages will immediately rise (and temps and fan too) as the gpu will now be eating about 30W at idle.
Note, the GPU will still clock lower if it hits power limit.

 

Thanks a ton! Registered just to chime in and say I had the same problem too. I'll try this soon, and if I can't get it working, I'll report back - I'd like to try as much as I possibly can before the RMA process, and I think this is probably the last straw.

Cheers guys!

 

I locked the voltage as per this chart, and I'm still getting some reboots in games - although it is a lot more stable now (I'm able to play Far Cry 5 for more than a split second). Anyone have any recommendations?

https://imgur.com/a/H9tpDdF

Nemesis90, are you still running stable on your machine? I'd rather avoid the RMA process if I can, but looks like I'm out of options.

(sorry, the insert image doesn't seem to be working)

 

You didn't do it right.
First, the curve is not fully flat.
Second, lock it and flatten it at the 0.900 or 0.950v voltage point, not past it.
Make all the points flat from there to the right.

 

Thanks for the help! How's this chart looking?

https://imgur.com/a/DAZwa3F

I don't really mess with Afterburner much, much less for purposes like this, so if I'm still messing it up, please let me know. Cheers!

 

Looks perfect. Should never crash now.

 

Thanks! I'll give it a go in just a few minutes here - if I disappear from the thread, it means all is well. If I still get reboots, I'll let you all know. Appreciate the help!