Driver Power State Failure - Vega 64?

[Trojan Horse]

My PC crashes on idle/load after half an hour of use
Sometimes audio gets microstutters, like via Spotfiy as well

Ryzen 3600
MSI B450 Tomahawk
32 GB RAM

I mix 2 sets of different ram, both E-Die I think (how do I check) but should be 2x8gb of Crucial AT Ballistix and another 2x8gb of Crucial LT?

Everything running factory, just updated the Vega and Tomahawk drivers

 

[Panos]

My PC crashes on idle/load after half an hour of use
Sometimes audio gets microstutters, like via Spotfiy as well

Ryzen 3600
MSI B450 Tomahawk
32 GB RAM

I mix 2 sets of different ram, both E-Die I think (how do I check) but should be 2x8gb of Crucial AT Ballistix and another 2x8gb of Crucial LT?

Everything running factory, just updated the Vega and Tomahawk drivers

What PSU you have?

 

[Trojan Horse]

What PSU you have?

Corsair AX860

 

[Panos]

Corsair AX860

The problem is not GPU related, but something else. Either CPU cooler (monitor the temps) or PSU.

 

[Tetras]

I mix 2 sets of different ram, both E-Die I think (how do I check) but should be 2x8gb of Crucial AT Ballistix and another 2x8gb of Crucial LT?

You could try Thaiphoon burner or Aida. If you're worried about compatibility just set them to a default JEDEC spec for awhile, preferably 2133. Though be aware you'll be overvolting them at that spec if you're running 1.35v.

p.s. are you running on an SSD? Did it always crash like this or start recent?

 

[Trojan Horse]

You could try Thaiphoon burner or Aida. If you're worried about compatibility just set them to a default JEDEC spec for awhile, preferably 2133. Though be aware you'll be overvolting them at that spec if you're running 1.35v.

p.s. are you running on an SSD? Did it always crash like this or start recent?

Yes I am running SSD

The problem is not GPU related, but something else. Either CPU cooler (monitor the temps) or PSU.

I think I have fixed it by setting Link state power management to 'maximum power saving' on Windows, after following a guide for the power state failure. I haven't had any crashes in 10h+ of continuous use but will have to test more before I can start OCing from Factory

EDIT: NVM it still crashes... Whats the best way to check if its temsp causing the issue?

 

[Panos]

Yes I am running SSD


I think I have fixed it by setting Link state power management to 'maximum power saving' on Windows, after following a guide for the power state failure. I haven't had any crashes in 10h+ of continuous use but will have to test more before I can start OCing from Factory

EDIT: NVM it still crashes... Whats the best way to check if its temsp causing the issue?

HWINFO64 Also remove one of the ram kits. Keep only 1 in on the dual channel slots.

 

[Tetras]

Yes I am running SSD


I think I have fixed it by setting Link state power management to 'maximum power saving' on Windows, after following a guide for the power state failure. I haven't had any crashes in 10h+ of continuous use but will have to test more before I can start OCing from Factory

EDIT: NVM it still crashes... Whats the best way to check if its temsp causing the issue?

Just to check, do you have automatic driver updates enabled? I don't know if it applies to Vega, but with RX cards, the 20 series drivers were a forced update for many people and they have crashes with black screens, freezes and audio loops. I reverted back to 19 drivers (19.11.3), disabled driver updates, like other RX users did and I haven't crashed in 4 months. Your problem might be different, I didn't get a driver power state failure, there was no notification, but could be worth a try.

If it is a power problem, could try underclocking/undervolting or swapping cables?

 

[Trojan Horse]

HWINFO64 Also remove one of the ram kits. Keep only 1 in on the dual channel slots.

Any way you can get temps in red for HWINFO? There is a lot of info and all the critical stuff is in its own tabs... hard to look at on my second monitor.

Which RAM kit should I take out first? AT or LT? (I am just gonna take out whichever is easiest anyway, as have massive K2 Mount Doom blocking everything)

Just to check, do you have automatic driver updates enabled? I don't know if it applies to Vega, but with RX cards, the 20 series drivers were a forced update for many people and they have crashes with black screens, freezes and audio loops. I reverted back to 19 drivers (19.11.3), disabled driver updates, like other RX users did and I haven't crashed in 4 months. Your problem might be different, I didn't get a driver power state failure, there was no notification, but could be worth a try.

If it is a power problem, could try underclocking/undervolting or swapping cables?

I don't know how it could be a power problem, as my system is not that heavy (anything I can screenshot to show you, but it is a 3600, vega 64, 3x 1tb ssd, 2x3tb hdd) although my PSU is nearly 10 years old now so that may be the case too!

Can you recommend any good ones?

 

[Panos]

Any way you can get temps in red for HWINFO? There is a lot of info and all the critical stuff is in its own tabs... hard to look at on my second monitor.

Which RAM kit should I take out first? AT or LT? (I am just gonna take out whichever is easiest anyway, as have massive K2 Mount Doom blocking everything)



I don't know how it could be a power problem, as my system is not that heavy (anything I can screenshot to show you, but it is a 3600, vega 64, 3x 1tb ssd, 2x3tb hdd) although my PSU is nearly 10 years old now so that may be the case too!

Can you recommend any good ones?

I hope you put the ram correctly. 1 kit on one dual channel the other to the next. And you didn't mixed them.
So I guess the second ram. Look your manual for the layout. It could be slots 1-3 (channel 2) and 2-4 (Channel 1)

 

[Tetras]

I don't know how it could be a power problem, as my system is not that heavy (anything I can screenshot to show you, but it is a 3600, vega 64, 3x 1tb ssd, 2x3tb hdd) although my PSU is nearly 10 years old now so that may be the case too!

Can you recommend any good ones?

I wouldn't have thought so either, but that's what the driver says isn't it? I do know that some people make their system stable by underclocking/undervolting their GPU, though that's not really a long-term solution. I guess you could try removing most of the HDDs and see if that lesser load helps? Like you though, I think it's unlikely.

I'm way too out of the loop to recommend a PSU unfortunately, I used to read jonnyguru back then.

If you have a spare graphics card that might be the easiest way to troubleshoot?

 

[Trojan Horse]

I wouldn't have thought so either, but that's what the driver says isn't it? I do know that some people make their system stable by underclocking/undervolting their GPU, though that's not really a long-term solution. I guess you could try removing most of the HDDs and see if that lesser load helps? Like you though, I think it's unlikely.

I'm way too out of the loop to recommend a PSU unfortunately, I used to read jonnyguru back then.

If you have a spare graphics card that might be the easiest way to troubleshoot?

Hey so ATM I don't have a spare GPU to trouble shoot. I could abuse the Rainforest's policy and return within 30 days... but don't know how ethical that is!

For the time being, I have found that putting my fans on max will keep me floating for an extra hour or so. This is strange because Furmark doesn't seem to trigger it? I think maybe something on the Tomahawk isn't cooling properly... I do have liquid metal on my 3600, so maybe the application wasn't done correctly?

How can I get alerted when any of my temps go in the red?

I hope you put the ram correctly. 1 kit on one dual channel the other to the next. And you didn't mixed them.
So I guess the second ram. Look your manual for the layout. It could be slots 1-3 (channel 2) and 2-4 (Channel 1)

As above

 

[hyperseven]

I assume this is a BSOD crash? If so have you parsed the dumps/minidumps?

 

[Trojan Horse]

I assume this is a BSOD crash? If so have you parsed the dumps/minidumps?

I haven't. This is good advice - can you recommend a guide to do so? Should I post them here as text?

Update:

Using BlueScreenView, 4/5 of my crashes are caused by the ntoskrnl.exe which suggests a memory problem?

sfc /scannow shows no issues
chkdsk /f /r running now

 

[hyperseven]

I haven't. This is good advice - can you recommend a guide to do so? Should I post them here as text?

Update:

Using BlueScreenView, 4/5 of my crashes are caused by the ntoskrnl.exe which suggests a memory problem?

sfc /scannow shows no issues
chkdsk /f /r running now

I always use this guide when finding the root cause of a BSOD https://www.tenforums.com/tutorials/5558-windbg-basics-debugging-crash-dumps-windows-10-a.html

 

[Holst1981]

All of my stability problems with the same board were due to my ram (reused from x99)

I would definitely look at the ram first.

Have you run memtest?

 

[Trojan Horse]

I always use this guide when finding the root cause of a BSOD https://www.tenforums.com/tutorials/5558-windbg-basics-debugging-crash-dumps-windows-10-a.html

This is a bit of an intensive guide... if my RAM checks do not work I will go back on to this when I have more time.

All of my stability problems with the same board were due to my ram (reused from x99)

I would definitely look at the ram first.

Have you run memtest?

Yep, I think it is definitely my RAM.

It turns out I am running Crucial Sport 16GBX2, maybe because I bought 2 seperate 16gb sticks instead of a 2x16gb kit ? specifically
BLS2K16G4D32AEST

should I run memtest overnight?

I have taken one stick out for now maybe this will fix things

Update: The BSOD dumps show an average uptime of less than two hours. With one stick removed I have been running 4hrs+ with no micro stutters (straight after boot I could hear Spotify stutter).

I will see about running memtest but I think this remaining stick is working. Do you guys know if Crucial will honour warranty during COVID?

 

[Holst1981]

This is a bit of an intensive guide... if my RAM checks do not work I will go back on to this when I have more time.



Yep, I think it is definitely my RAM.

It turns out I am running Crucial Sport 16GBX2, maybe because I bought 2 seperate 16gb sticks instead of a 2x16gb kit ? specifically
BLS2K16G4D32AEST

should I run memtest overnight?

I have taken one stick out for now maybe this will fix things

Update: The BSOD dumps show an average uptime of less than two hours. With one stick removed I have been running 4hrs+ with no micro stutters (straight after boot I could hear Spotify stutter).

I will see about running memtest but I think this remaining stick is working. Do you guys know if Crucial will honour warranty during COVID?

Two sticks are "harder" for the CPU to run than a single stick.

Test both sticks individually

You will needs to tune voltage and timings to get both sticks working together.

 

[Trojan Horse]

Two sticks are "harder" for the CPU to run than a single stick.

Test both sticks individually

You will needs to tune voltage and timings to get both sticks working together.

OK so MemTest86 came by with a pass

Except running my PC on Idle, with fans turned off, caused same BSOD error... my K2 Mount Doom is passively cooled enough to run the Ryzen chip so I don't see how the temp causes it this time