MSI R9 390x drivers crashing

[Jodz00]

I'm fully aware that multiple threads have been made which highlight this issue but I am still without a solution. When I first got my card, I had a problem with the drivers crashing my PC; this was quickly fixed by upgrading the drivers. Recently, I've been getting frequent pop-ups letting me know there is a new version of my drivers. I didn't want to upgrade at first, in fear that the crashing may reoccur, so I put it off until yesterday. To no surprise, my drivers have been crashing again. I've tried multiple different drivers but yet to find any success - am I missing something?

 

[RaDiKaL_]

If you're having crashing issues with different driver versions then there's probably some hardware issue going on. Stress test your GPU with Furmark and monitor temps and voltages, you can use HWInfo for this.

Full pc specs and specific instances when it happens could also help to clarify the situation.

To ensure no driver issues uninstall once more using DDU utility and then install the driver you'd prefer to use.

 

[Jodz00]

Apologies for the late response but I carried out a stress test and the PC did crash. It seems odd that only now I'm experiencing these hardware issues as the card has been fine.

 

[RaDiKaL_]

Which PSU do you have?

 

[Jodz00]

An Aero Cool GT-700S

 

[RaDiKaL_]

Well it's not the greatest but should deliver enough for your GPU, HWInfo log should give us the entire scene of what's going on. Go ahead and take a look at its log, look for the GPU usage/temp and voltages while running the test and compare it to its last data right before the crash.

 

[Jodz00]

Would it be easier if you took a look at the logs yourself? http://www72.zippyshare.com/v/8CJs7DJF/file.html

 

[Jodz00]

Just an update: I've been monitoring the temps of the gpu while attempting to play games and it's been fine. When the pixels come up before the crash, there is little change in temperature.

 

[RaDiKaL_]

Wow... it was really interesting to see even though the data recorded was just a few minutes.

So, first of all, before and after running Furmark, there seems to be something using your HDD at 100% both reading and writing, from the logs I can tell you it's not the pagefile nor some hefty download/upload in your network-internet, so that's something you need to find out using procexp.

Now into the Furmark test, your GPU reaches 92ºC+ and before the crash it gets up to 3872253 memory errors, that's obviously causing the crash, the errors start to accumulate when your GPU reaches 89ºC which coincides with a current spike of 258W into the GPU.

To this I must add that your PSU's 12V+ line drops to 11.737v and in the GPU to 11.313v, both values are not good, it shouldn't drop below 11.8xxv.

You need to solve the temp situation first, the instability that it causes could be the reason the 12V+ line is dropping below expected values, seeing some reviews your GPU should never go above 80ºC. If after fixing the temp the 12v+ still drops that much then you'll need to consider getting a better PSU, your CPU is also getting affected causing two of your six cores dropping from time to time from their 4.2ghz (4118mhz in log) to 1.7ghz to get immediately back to 4,2ghz, your CPU temps are fine so this is the result of power instability.

Make sure your GPU fans are working and if necessary you'll need to replace its thermal paste.

 

[Jodz00]

I changed the gpu fan settings to be constantly running at a faster speed to keep the temps down; has crashed less since. Will it be necessary for me to replace the psu?

 

[RaDiKaL_]

"has crashed less since", with the new adjustment its necessary to see the new temps your GPU is reaching, if it's below 80ºC and still crashes then yes, the PSU would be the next one in line.

 

[Jodz00]

I've purchased a new psu (rm750x) but I seem to be experiencing some issues with that. I will need to resolve to first, in order to find out if the initial crashing problem has been resolved.

 

[Jodz00]

New psu installed with a new hard drive, it killed the old one for some reason, but the problem still remains.

 

[RaDiKaL_]

As said before, we need to see the temps with the changes you've made, maybe ramping up your fans is only making it longer to reach above 80ºC.