Kernel-Power Critical error in Win 10

[Raswan]

Hey everyone,

So as you can see this has happened ten times over the last 6 weeks. Google-searching has led me nowhere so far. Can anyone point me in the direction of more info about what's causing this error and how to fix it?

https://www.dropbox.com/s/leaodt6bxey9tb8/kernel-power issue.JPG?dl=0

 

[UsandThem]

Have you tried turned off 'Fast Startup' in Windows?

https://www.howtogeek.com/243901/the-pros-and-cons-of-windows-10s-fast-startup-mode/

However, there are a lot of people with that event showing up in logs. It really doesn't hurt anything if you are able to simply ignore it.

 

[deustroop]

You haven't said what you are doing on the computer at the times noted nor what the error message on the desktop might be. There's a whole part of this thing that is missing, eh ? The events are described as critical so we need more info.

 

[mikeymikec]

If it's a problem with fast startup then there's usually another event log entry nearby saying something along the lines of 'fast startup failed with status of ....'.

But it doesn't hurt to switch it off for a week and see what happens.

Windows 10 has an annoying tendency to post a lot of crap in the event logs as errors even though for example a service is working according to design.

 

[Ketchup]

Is the computer losing power? Or does it just say it is?

 

[ch33zw1z]

Definitely need more details.

What type of computer?
Desktop, laptop?

Hardware specs?

Firmware up to date?
Software up to date?

 

[Raswan]

Hey all,

Thanks for the replies. I should have provided more info. Apologies. Usually when I post a thick paragraph the help thread just gets ignored. Here we go, let me know if you want more info that I'm missing.

This is a desktop PC with an i5-6500, 32GB of DDR4, Asus Z170a mobo, 850 watt EVGA platinum power supply.

So I initially thought these issues were GPU-related. I'd received an R9-390 as a replacement for a 280x from XFX earlier this year and have been having problems for the last month or so. Mostly, I'm crashing out of games (Overwatch, Paragon, Rocket League) and (much less often) Windows is locking up, with a display driver failure error given. I have uninstalled and reinstalled video drivers a few times (both using a driver cleaner and not), which seemed to fix the issue temporarily, though not this most recent time. Today it simply locked up, displaying a bunch of artifacts, then BSoD'd. Restarting it would only get me as far as the login screen, and then the process would repeat. My suspicion was that the card was overheating--I've been tracking temps in low-intensity games like Rocket League and Paragon, and as soon as I launch the game the fan winds up to about 90% and the temp hits 90-92C within a minute or two (I know 90c is nothing to worry about in general, though it's at the top end of the operating threshold. Just seemed off to me that in a game like Rocket League the card is hitting those temps within five minutes). One particular night when Paragon kept crashing I took the side of my case off and that seemed to fix it. In Windows temps are fine and much less crashing happens. Note that my InWin 904 PLUS is clean of dust, has good airflow, and my CPU sits at a comfortable 34-8C at idle and maybe 50C at load. I use all Noctua-brand fans and turn them up when gaming.

Replaced the 390x with a 5870 I had around, cleaned and replaced drivers, and everything seemed cool. Then got the BSoD again. This was two days ago, no crashes since.

Two other issues:
1) While playing Rocket League, if I try to alt-tab, Windows locks up. It restarts after about ten seconds, everything being fine after. But if I alt-tab again, the exact same thing happens.
2) If I try to listen to youtube while playing, chrome has a fucking fit and locks up hardcore. Windows recovers if I end chrome in the task manager, but chrome remains unusable until a restart. This is the only time chrome gives me trouble.

Ideas? I am no longer sure it's the GPU with the most recent crash and in looking at the event logs initially, though it sounds like Windows can be dramatic in logging critical errors. I will say each one of those critical errors in the linked image coincides with a complete lock-up and a hard restart.

Ideas?

Mikey: I'll disable fast startup for a couple weeks. Good idea.

Ketchup: Just losing display, computer keeps running.

Ch33z: Only thing I'd forgotten to update was BIOS. Did that today, but again, haven't had any crashes since the 5870 was put in there. This doesn't necessarily mean anything, since I'd go a week or two without one of those crashes.

Thanks guys. At the end of my rope a bit here.

 

[mikeymikec]

It seems to me that the event log messages you're talking about are the least of your problems. That CPU temp does not sound right even for the stock cooler, especially as that's a non-K CPU and so therefore shouldn't be overclocked (I'd double-check it isn't though).

Memtest86+ overnight (non-SMP)?

 

[Raswan]

It seems to me that the event log messages you're talking about are the least of your problems. That CPU temp does not sound right even for the stock cooler, especially as that's a non-K CPU and so therefore shouldn't be overclocked (I'd double-check it isn't though).

Memtest86+ overnight (non-SMP)?

What seems off about the CPU? Those temps seem fine to me. Looks like I was a bit under there. CAM shows it at 38C right now at idle, low of around 34 in the last hour. Let me head into a game quick and see what it hits at load.

Edit: Looks like it hits a max of 50C under load.

The cooler isn't stock. It's this Noctua.

 

[mikeymikec]

What seems off about the CPU? Looks like I was a bit under there. CAM shows it at 40C right now at idle. Let me head into a game quick and see what it hits at load.

What's the environment temperature? I'd expect a Skylake i5 in my house to be about 30C when idle, not 40C.

Don't use a game for testing CPU load temps, use Prime95 and push the CPU to its limits.

 

[UsandThem]

Yeah, try a utility like Prime95 (I personally like Intel Extreme Tuning Utility better though). But if you are using the stock i5-6500 cooler, and are at 30c and 50c load, that's nothing to worry about.

However, the R9-390 hitting 90c+ that quickly is not good at all. It seems most people are in the mid-to-upper 70s with that GPU.

But since you have had issues with two XFX cards in a row, the issue could be PSU and motherboard PCIE X16 slot issue (unless you are just unlucky with video cards).

 

[Raswan]

What's the environment temperature? I'd expect a Skylake i5 in my house to be about 30C when idle, not 40C.

Don't use a game for testing CPU load temps, use Prime95 and push the CPU to its limits.

Ambient temp in this room is around 75 Fahrenheit. Never gets above 80. Warm, I know, but not crazy.

Good point about the stress-testing. Can't remember if I have prime 95 installed. I'll dl and run a cycle and see what happens.

 

[Raswan]

Yeah, try a utility like Prime95 (I personally like Intel Extreme Tuning Utility better though). But if you are using the stock i5-6500 cooler, and are at 30c and 50c load, that's nothing to worry about.

However, the R9-390 hitting 90c+ that quickly is not good at all. It seems most people are in the mid-to-upper 70s with that GPU.

But since you have had issues with two XFX cards in a row, the issue could be PSU and motherboard PCIE X16 slot issue (unless you are just unlucky with video cards).

I'd had a thought that XFX sent me a refurb, which in my mind increases the chances of getting a dud. But I agree, two bad ones in a row is generally unlikely. Mid-70s is exactly what I'm getting with this 5870, and what I would expect to see from a modern-gen card as well.

Between today and tomorrow I'll reinstall the 390, update the BIOS for this mobo, and run prime 95 and memtest and post results. Thanks for hanging with me so far guys. Much appreciated.

 

[mikeymikec]

Ambient temp in this room is around 75 Fahrenheit. Never gets above 80. Warm, I know, but not crazy.

Good point about the stress-testing. Can't remember if I have prime 95 installed. I'll dl and run a cycle and see what happens.

I normally run Prime95 for half an hour, fyi.

 

[UsandThem]

I'd had a thought that XFX sent me a refurb, which in my mind increases the chances of getting a dud. But I agree, two bad ones in a row is generally unlikely. Mid-70s is exactly what I'm getting with this 5870, and what I would expect to see from a modern-gen card as well.

Between today and tomorrow I'll reinstall the 390, update the BIOS for this mobo, and run prime 95 and memtest and post results. Thanks for hanging with me so far guys. Much appreciated.

I wouldn't put it past XFX sending a refurb card with issues. Several years back I owned a XFX 7950 card that began checker-boarding my screen while at the desktop at random times (like in Chrome). Like you, I threw in an old AMD card to see if I could figure out the issue, and it worked just fine. When XFX got the 7950 card in I sent, they told me there was nothing wrong with it, but they were going to send me a new replacement for it anyways (which I found odd). Anyways, when the replacement card came in, I put it in and it didn't have any issues for several years until I sold it. So the first card I sent in did have issues.

 

[Ketchup]

The messages in your initial post are just showing your hard reboot, so they are not directly related to the issue. Have you ever used CPUID hardware monitor? Run it while you are launching a game and try to close the game before it crashes the computer, but after fans have hit top speed. Send us a screenshot. What we need to be looking at in min and max temps for CPU cores (they are all listed separately) and GPU.

 

[ch33zw1z]

The GPU stands out in my mind from all this.

The cpu temp looks fine

My i5 6600k idles about 31 C, converting video in handbrake bumps it to high 50's

Well within specs

 

[Raswan]

I normally run Prime95 for half an hour, fyi.

Will do. Thanks

 

[Raswan]

The GPU stands out in my mind from all this.

The cpu temp looks fine

My i5 6600k idles about 31 C, converting video in handbrake bumps it to high 50's

Well within specs

Same here. Haven't seen any CPU spikes or generally high temps that suggest the cooler needs to be re-situated or thermal grease re-applied.

 

[Raswan]

I wouldn't put it past XFX sending a refurb card with issues. Several years back I owned a XFX 7950 card that began checker-boarding my screen while at the desktop at random times (like in Chrome). Like you, I threw in an old AMD card to see if I could figure out the issue, and it worked just fine. When XFX got the 7950 card in I sent, they told me there was nothing wrong with it, but they were going to send me a new replacement for it anyways (which I found odd). Anyways, when the replacement card came in, I put it in and it didn't have any issues for several years until I sold it. So the first card I sent in did have issues.

Good to know. The 280x this 390 is replacing had a slew of issues. Sent it in, they claimed they couldn't replicate and sent this 390 "anyway" as a replacement. So sounds like we had the same situation.

 

[Raswan]

Well, got one of these bad boys last night, so I guess now I'm leaning towards mobo instead of GPU. I'm out of town for the weekend, but still intend to prime 95 and memtest when I return. Any other things I can do on top of that to isolate the mobo as the problem? It's only 16 months old, so definitely still under warranty.

 

[ch33zw1z]

What type of p95 testing were you doing.

See how the screen is all messed up, bet it's still the vga, lol

 

[UsandThem]

You can test the video card in another PCIE slot to rule out the slot being the issue.

One thing I am starting wonder about is your PSU. I know it's a nice EVGA Platinum unit, but maybe there is an issue with it, and it's over-volting your GPU. Maybe take a look and make sure the cables are fully seated, and plug them into different VGA ports (your unit has VGA 1-4). I think your unit comes with 4 PCIE connectors (2 x 6 + 2 pin) and (2 x daisy chain), so maybe try using different cables.

 

[ch33zw1z]

Definitely a good point.

The bios will likely give you readouts for each voltage. Make sure they're within +/-10-15%

 

[UsandThem]

Also, try running this and see if it sheds any additional info on your bluescreens:

http://www.resplendence.com/whocrashed