help needed: how to determine if GPU is dying?

Here is what happens:

When I run games, the laptop crashes about half the time.
It can be Lego Starways from Steam, or Overwatch.

Laptop: aorus x5v6
OS: Windows 10 april 2018
Drivers: Nividia 12/12/2018 standard drivers 417.35
Temps: I have seen some 90 degree temps when playing overwatch.
Type of crash: when playing overwatch I get two kinds of crashes: first kind is overwatch crashes and error message says "rendering device lost". other kind of crash is the entire laptop reboots. This happens when going into the game itself, or going into the video options in overwatch. Note as above, this also happens in other games such as lego starwars.

I have restored the system from scratch, same problem.

I ran furmark for 5 minutes and no crashes.

What would be the best way to determine if my gpu is dying?
I have already googled and get a lot of crap answers from toms hardware and that sort of stuff, so though I would post here in nbr and see if I can get a knowledgable answer on what kind of software to run, what logs to look at, and how to tell for sure what the problem is.

Thanks

 

download Hwinfo64, run logging and play a game until it crashes, then send the log here.

 

Thanks here are those log files, right after overwatch crashed:
Sensors CSV log: https://docs.google.com/spreadsheets/d/1RWLjgC1dJTrS0y4ecDlIHHrmObXY46NCZTin7J9Z_Ew/edit?usp=sharing
System snapshot log: https://docs.google.com/document/d/1uBP_nmsm8Q3qboZRTeg8zzSkT0r1WpENNyWO991QEtI/edit?usp=sharing

 

You are not throttling cpu temps look good and gpu reached 75C before reducing.

Everything looks normal.
Hard to pinpoint this one. COuld be psu related could be motherboard.
Will not be easy to identify.

 

here is shorter log file from today with less info in the sensor logs, mainly just gpu stuff ( not hard drive sensors and no basic cpu sensors ) it is much shorter.
however I don't see anything that indicates an issue:
here it is: https://docs.google.com/spreadsheets/d/1jhzQ78F8-wYQoiMecleXZWvv5FRIr7wioR_AXdTAJYk/edit?usp=sharing

 

thanks. I have a kill-a-watt on my input to the power supply and it is not going over 180W

 

is this while you played overwatch?

 

yes.

 

Have you tried going back to a non-400 series driver?

 

Can you remove the kill-a-watt and see if the problem persists?
Do you have another psu to test?
Trying to identify hardware or software.
We going with hardware first.

 

Yeah so I am absolutely confused as to why nobody sees the issue here.

Your CPU runs abysmal, sub 1ghz clocks in game

 

does that low clock rate mean there is a root cause that leads to the crashes, and what do you recommend I do to take action to fix this?

 

it happens without the kill-a-watt I put that in place to see if it was drawing too much ( but it is not ).
is there any way to get a DC power reading from the laptop to see what the power supply is supplying? I don't want to spent $75 on a new power supply if that is not the problem

 

Not the neccessary the root, but casued by the root of the problem. You might have a failing power supply or something is wrong with your mainboard.

 

any suggestion how to diagnose power supply?

 

Trying another preferbly.

 

OK I'll think about that.
thank you for the troubleshooting help.

 

EDIT: What are the chances something about xtu has caused this? let me tell you why I ask that. I on a whim uninstalled xtu and command and control, and was unable to create the issue again. then I re-installed command and control, and the issue came back. then I uninstalled c&c, and could not recreate issue. then I installed xtu, and the issue came back. is this just a coincidence or is there something to it? I wish I could controlt he fans without c&c but speedfan does not recognize this aorus/gigabyte fan

 

So the CPU runs at normal speeds and no crashed when you removed the stuff? If so then there is your answer. Maybe look for another version of the control center.

 

thanks. that's fairly obvious even someone like me knows that.
I have tried other versions.
I will post back to the thread if the problem comes back.
I will buy another p/s today.
Do you think I should get the same 10.3A 19.5V which is the same it comes with, or go bigger to an 11.8A 19.5? The bigger one is not going to fit in laptop bag, and is $30 more, it is a large square brick.

 

I would ask : did you crash because of your undervolt or whatever the settings you were changing with XTU, or was it just simple fact of running XTU in the background, cause you the problem? Because normally XTU will reset all settings to original after the crash and it will be the very first thing to look at, to find a cause of crash, but sometimes I had issues even if XTU was just running in the background with stock settings. Also, keep in mind XTU has a driver that loads on bootup, regardless if XTU itself is ever run. You need stop the driver from loading in task manager or remove XTU altogether.

 

Hi sorry for late reply, I think it was just because XTU was running in the background. It's pretty stable now because xtu no longer runs and "command and control" no longer runs, uninstalled both. - I have had one crash in last 20 or 30 overwatch sessions see log below ( short log this time! ) but not like every 3rd time playing like what was happening before, so removing xtu has made it almost totally better. Here is latest hwinfo64 sensors log with ALL sensors on, there was a full system crash while playing overwatch near the end of this, had to cold boot to fix it, here is the short log file: https://docs.google.com/spreadsheets/d/19rXfjX6hL8BpivtYI9p1gJ0yevdOEYfb0ovPqjz3YoI/edit?usp=sharing

 

So it was software related not hardware.
Moral of the story?
Tl: /DR Always remove rogue software before testing.

 

Who would have thought XTU from Intel would crash the system even on stock settings. I had few unexpected shutdowns and at first I didn't believe it was caused by xtu itself, I even double checked the voltage settings, thinking, maybe I forgot and somehow set it too low, but it was just running in the background at stock and since I removed it, all is fine. I also had issues with speedfan. Bottom line is, with so many different hardware combinations it's almost impossible for software makers to test every configuration and pretty much any program can be a suspect and no, I would not consider XTU as rogue.
So I would say moral of the story is: don't assume anything and run your system as lean as possible. BTW on my system Throttlestop works much better than XTU.

 

I believe Aorus has made mention (or I've read elsewhere) that XTU and C&C don't play well together. C&C has overclocking utilities that can create problems being layered on top of XTU. Additionally, and this is from my own experience, I'm not impressed with quality control at Gigabyte/Aorus. Because you had that one crash even after uninstalling XTU/C&C, I'm not convinced this is entirely a software issue.

After getting rid of my X5V7, the alienware that replaced it has been crash free for more than a year. My G75VX before it has never crashed in the 7 years I've owned it (still chooching, its my kid's rig now).

 

update: spent $55 and got a big heavy 230W power supply brick. Now my kill-a-watt shows up to 225 watts being drawn when playing overwatch. so far no crashes. will keep testing. but if the laptop is sucking 225W, then the 200W power supply that shipped with the laptop would certainly not have delivered the required power. Based on the original diagnosis by some nice person to my first couple of posts, that certainly seems like the right direction to go in, so I'll keep testing and if I have any more crashes I will reply here, if I don't, it means there was simply not enough power.
slight edit: I wonder why then I was able to play for years without issue? Is my laptop now drawing more power than it previously did? if so why? I am not running xtu service. could there be some kind of residual setting from when XTU was running that is stuck in the system, causing more power draw?

 

update: even with plenty of power coming in, still get crashes. now playing apex legends, still get them. xtu installed or not, still get the crashes. using afterburner to max underclock gpu and gram makes it only crash once ever hour of playing, if I don't use afterburner, it crashes every 10 minutes.

 

Bump. Is there an FAQ I should be reading hidden somewhere on this site that would help guide me to understand if my hardware is dying?
Based on this thread, does anyone think my GPU, Mobo or something else is the issue?

 

I wanted to reply to this thread in case it helps anyone else who was having the same problem as me.
I don't think buying a larger power supply had anything to do with it, as that did not fix the problem.
However, within a few hours of posting on the x5v6 forum, it was suggested to me to ctrl-F and then shift-drag the curve down in afterburner to get it lower than what the sliders do.
I went down to -700 on the clock and -500 on the memory, and no more crashes.
hope this helps someone.
-bw