Computer freezes with installed graphics card drivers, not without

[swiggu]

Title.

I get both freezes where the screen turns black, my monitor goes on standby and there's a couple of seconds of buzzing coming through my headphones and freezes where the screen turns still. In both cases I get no error messages whatsoever and my only option is to manually shut down my PC with the power button or reboot it with the reboot button. Finally, I sometimes also but much less frequently get the message that the AMD display driver has stopped responding and has recovered, where I get a few seconds of unresponsiveness but it comes back up after a while.

I previously thought the issue was lying in certain Microsoft/Windows services, specifically  Superfetch and Windows Image Acquisition, as I didn't get the freezes for a couple of days with them disabled on startup, but the freezes came back after a while nonetheless.

I tried to remove the drivers in the Device Manager and I haven't had a freeze until now, but as I like to game and use Netflix and the likes, I feel like using my actual graphic card's drivers would be optimal for displaying performances, so I'd like to find the root of my issue. I don't seem to get the freezes when using the standard VGA graphics/monitor.

It might have something to do with the voltage of some of my hardware, but I'm not at all accustomed to making changes with stuff like that, so I might need some guidance if that's a possible solution.

I think I've downloaded every driver and driver update for the rest of my hardware using drivers supplied by the manufacturers and their websites respectively, so I doubt it's related to out of date drivers, although it could be some sort of hardware/driver conflict between my graphics card and any other device, who knows.

Here's my specs:
- Windows 7 Home Premium 64-bit SP1
- AMD FX-6300, Vishera 32nm Technology
- 4,00GB Dual-Channel DDR3 @ 799MHz
- MSI 970A-G43 (MS-7693) (CPU 1)
- Standard Monitor (otherwise Club3D Radeon R7 260X royalKing Vapor with drivers installed)
- 931GB Seagate ST1000DM003-1CH162 ATA Device (SATA)
- HL-DT-ST DVDRAM GH24NSB0 ATA Device
- Realtek High Definition Audio

Thanks in advance.

[Calum]

Was this a working system you added the card to, or is it a new build?

[swiggu]

It's a new build.

I should note that I've ran Memtest86+ a couple of times, so I think RAM is also excludable as a cause. I'd put this in the first post but I can't find the edit function.

[Calum]

Edit is disabled after the post has been up a while, so putting the info in a new post is fine

Which AMD driver versions have you tried so far?  Is there any pattern to the freezes, for example do they happen only when (or are they more frequent when) running Flash video or web browsers, or anything like that?  It could be worth disabling hardware acceleration in Flash and your browser(s) to see if this helps at all.

[swiggu]

So far it can generally happen at any time I'm actively using my PC for around 20 minutes, but most frequently when I'm running things in full screen, like videos, be it in Chrome or in VLC, or video games, but I've also had it happen when I was just writing a post or looking around the services.

I've tried a couple of different ones, but none of them to even install for me except the most recent 13.11 Beta 9.4 version.

I turned on OC Genie II in my BIOS purely out of curiosity, and I haven't had any of the irrecoverable freezes for a few hours now, but I did still get the AMD display driver crash which recovers after about half a minute, but forces me to reboot anything utilizing my GPU for rendering which is annoying.

[Calum]

Yeah, that figures with it being almost a brand new card (well, in name & device ID anyway).
I don't see a PSU listed in your specs, what make & model of PSU are you using?
Could also be worth trying the card in a different physical PCI-E slot, shouldn't make any difference but it's worth a quick try.

[swiggu]

My PSU is a Sharkoon WPM 500W.

I could try that out over the weekend when I have access to all of my tools (I live on campus and my strap and screwdrivers, etc. are all back home).

[swiggu]

Is there anything else I could try out in the meantime?

[Calum]

If you have a different PCI-E connector on that unit, try that - again, it shouldn't make a difference, but it's worth a try.
In a case like this where you have a brand new card malfunctioning, on a brand new system, with the crashes seemingly related to the card alone - you have limited options really.  Different drivers aren't an option as the 260X is new (as I said, in terms of its name and device ID although it's basically a 7790), so other than trying a different slot or PSU conenctor, I can't really suggest much else other than returning the card.

[swiggu]

Even though I have an additional PCIe slot on my motherboard, I can't switch it since the connectors for my hard drive are in the way and my wireless network card is occupying the slot.

I also checked the event manager and when I got screen resets, it showed error ID 4101 and that "amdkmdap had stopped responding and has recovered", and when I got complete freezes it showed error ID 41, which was kernel-power related. Could it be a combination of the AMD driver malfunctioning in combination with an error with my kernel, or just an AMD driver issue?

[Calum]

Kernel power event 41 just indicates the machine hasn't shut down properly, you'll often see the same event after a power cut or if you pull the plug.  It doesn't refer to the OS kernel in this case.
You're basically dealing with a graphics card not functioning correctly, as I said this can be down to the card itself, the PCI-E slot, PSU or connectors, or the driver.  There's very little you can rule out with there being no other drivers available, and the slot and power cables are outside chances at best, so I do suspect your card needs replacing.

[swiggu]

Alright. Would you suggest I replace my current card with the same model or a different one?

[Calum]

It's entirely up to you, if that card matches your requirements then by all means just RMA it for a replacement, if you'd prefer something different then RMA it for a refund or credit and get something else.
I can't comment on the overall reliability of the 260X as it's too new for any meaningful stats, so if you were asking due to reliability nobody can really answer that yet.

[swiggu]

Alright, thanks. I'll just get a replacement for the same GPU then.

[swiggu]

I tried reseating everything and got one regular AMD display driver crash into a recovery, then an hour later I got an actual BSoD for the first time that created a mini dump. Does this change my situation at all? Are there solutions for this?

[Calum]

I don't think it changes the situation, no.  I still suspect the card is faulty.

[swiggu]

Alright, so after sending it back they just sent  me a reply telling me they tested my card for multiple hours under load of video games, videos, etc. on a system of theirs and they're sending it back, so it could potentially be an issue with my motherboard and/or its PCIe slot. Thoughts?

[swiggu]

Alright, so I bought a new graphics card, this time an ASUS nVidia GeForce GTX 660 because I suspected drivers to be the cause of my issues with my other card, but even with this new card I got another system freeze. Could it potentially be that my motherboard or CPU are the culprit here?

[swiggu]

bump?

[swiggu]

Reposting this from another thread I made, but a mod closed it and suggested I only use this (even though it's pretty dead right now):

"So, around a month or two ago I bought and assembled a new PC.

Right off the bat, I've been getting regular freezes after installing all the drivers, notably those of my video card.

At first I thought my video card was faulty, and soon purchased a new one from a more reliable manufacturer (ASUS, previously Club 3D). Although this did reduce the frequency of these freezes, they still occur, and in the same way. As of now, I think this can outrule it being by video card.

I've ran multiple passes of RAM tests with Memtest86+, ran benchmarks on my CPU (Prime95) and video card (Furmark) to no live repetition of the problem for about 15 minutes for each (the amount of time recommended for both on other troubleshoots I've come across). In these scenarios, these units of hardware should be stressed to the max and thus forcing the freeze to reoccur, but it doesn't. As such, it may be my fairly low-end Sharkoon 500W WPM PSU that could be the problem, as I've heard it may sometimes provide tens of watts less in power (around 450~, whilst advertising 500), as I've noticed a critical error called "Event ID 41: Kernel power" in my windows event viewer.

It is either the aforementioned or my CPU (AMD FX-6300) at this point, perhaps a general lack of certain parts of my PC.

I really do think I'm past my options if it's none of the above and I'd appreciate any help I can get!"

[Calum]

Event ID 41 indicates the PC was shut down or rebooted without going through the normal procedure, you'd see the same if you'd had a power cut, hit the reset button or yanked the power cable out.  It doesn't necessarily mean the problem is with the power.
It could be your PSU, yes, but if Furmark and Prime95 don't cause this issue, it would seem unlikely, as they would of course cause the PC to draw more power.  The WPM 500 is a decent enough CWT based PSU, similar to lower end Corsair units for example, and it's definitely not something I would write off as incapable of powering your PC reliably.
Similarly, faulty CPUs are very rare, so I wouldn't like to point to your CPU as the problem here.
Have you been able to try a different slot or power connector as I suggested earlier, just to rule them out?
If you do still suspect your PSU then I would perhaps suggest trying to borrow a known good unit before buying a replacement.

[swiggu]

Ah, thanks, didn't know that.

Yeah, it's weird. It seems to happen when I'm using Chrome for some reason, specifically when using Flash-based stuff. I switched to Pale Moon and it hasn't happened yet. Could just be a coincidence like the myriad of other things I thought would be conclusive.

I've tried, yes, but my PSU connector is hardwired into the PSU itself, so there's no way of using another one unfortunately. It would also be weird if the connector just worked all the time, and then under stress would randomly be faulty. Doesn't sound logical to me, but I'd definitely try it out if I could.

I'll try borrowing a friend's PSU and try it out from there.

[patio]

All PSU units are made that way...
I believe Calum meant to try another 4/6 pin connector to the vid card...not switch or replace it at the PSU end...
That being stated borrowing a known good PSU of the same or greater wattage is the best thing to try at this point.

[Calum]

Per the data sheet here it looks like you have a 6 pin PCI-E power cable and a 6+2 pin PCI-E power cable, hence why I suggested trying the other PCI-E power connector.  It's a long shot, but worth a try.

Have you tried disabling hardware acceleration in Flash and Chrome/Pale Moon, to see if this helps at all or indeed makes it worse?  To me things are still pointing towards an issue with your graphics although getting two faulty cards in a row would be unlikely - again, hence why we're looking at your PSU as a possible cause.

[swiggu]

So, you're suggesting I insert only the 6-pin part of 6-2 into the card? Gonna try it out.

[Calum]

Correct.  The connectors are identical, the additional 2 pins are extra ground pins for 8 pin PCI-E power sockets.

[swiggu]

Alright, it's done. The other connector was seperate though (had to connect one end to the PSU and the other to the video card). It's too soon to say the problem's gone so I'll tell you if it happens again or not. Thanks for the help!

[swiggu]

Aaaaaaaand it happened. At this point it's either my motherboard or PSU... Thinking PSU, as it doesn't happen a lot when I use my PC after a night, but then after a couple of hours it does, and rebooting it makes it happen more frequently. Could be a heat problem somewhere as well. Ugh, this is so frustrating... 

[Calum]

Yeah, it was a bit of a long shot, but worth ruling it out.
Pretty much time to try a known good PSU if you can, see if that helps.

[swiggu]

Alright. Any PSUs you could recommend? I've asked some friends if they could bring over their PSUs (450 watts), but they were very reluctant and didn't want to risk breaking them or anything, even though I don't think there's any risks.

[Calum]

It's almost impossible that your machine could damage another PSU unless you somehow shorted it, but I appreciate not everyone wants to lend a PSU.  If you're in the UK you're welcome to borrow one of mine - otherwise, let us know whereabouts in the world you are, and we'll recommend some models.

[swiggu]

I live in the Netherlands, unfortunately, so I think I'll just have to purchase a new one. I could also just return my current PSU to the retailer I ordered it from (Alternate) and have it tested there, much like I did with my video card.

[patio]

That sounds like your best next option then...

[Calum]

Because faults like this don't always show up when testing a PSU unless they have some pretty expensive testing equipment (or your system) I would probably go with buying a new PSU.
Looking through the Alternate site (I'm on their UK site because I can't be sure of what I'm reading on the NL site, hopefully they stock the same products) the XFX Pro 550W, Silverstone ST50F-ES, and Antec Neo Eco 520 stand out as good quality, good value for money units.

[swiggu]

Alright, going for the XFX PRO550W Core Edition then. Fingers crossed...!

[patio]

If you are buying from the same shop you got the card from have them hook everything up and run some stress tests on it before you hand them your money...

[swiggu]

Alright, so a week ago I purchased that XFX PSU and replaced it with my old Sharkoon PSU. After almost a week of not having any freezes, it just now happened again. And I was about to come back here tomorrow and post that my problem was solved, too...

I'm really wondering if this is gonna be a one time thing or if it's gonna reoccur. Since I installed my new PSU I've noticed that, even though it didn't freeze, I got these rather unresponsive hangs (especially in Firefox when loading pages or in-game when playing DotA).

Either way, is there anything else I could potentially try out?

[Calum]

Have you already disabled hardware acceleration in Firefox?  If not, do that, it's never worked too well for me and has sometimes caused freezes.
Realistically the only other component that could really be at fault is your motherboard but it's very unlikely to cause a problem like this.
As the 13.12 drivers have been released since your original post, have you tried them, to see if the newer version helps with this issue?

[swiggu]

I haven't. I just did it. I'm not sure if it's gonna fix my system, however, as it occured when I was playing DotA. Firefox was open and running, however, although no video or page utilizing Flash was active at that time if I recall correctly.

I haven't. I bought a new video card, remember (nVidia GTX 660)?

[Calum]

Apologies...I missed that part.
I know the crashes occurred in game as well, I was just checking if it helped with Firefox at all.  Firefox's hardware acceleration is separate to the acceleration in Flash.
Try setting the power management mode in your Nvidia drivers to maximum performance - that's under manage 3d settings in the Nvidia control panel.

[swiggu]

No problem, I'm more than happy and grateful that there's people willing to help me out here haha.

Done. I'll see if I get freezes again now.

[swiggu]

It froze still. Running Speedfan and making it log to see if it's a computer part overheating right now...

[swiggu]

Alright, so the freeze happened again with SpeedFan logging temperatures and voltages, etc. Could anyone help me check out if there's anything wrong with my temps in this? Log attached.

[recovering disk space, attachment deleted by admin]

[Calum]

Well, your temperatures look fine to me too, no problems there.  You can safely ignore voltage readings as Speedfan and other software won't read, for example, your 12V voltage properly.
Is there anything in your event logs, other than that the driver has stopped responding & recovered?

Seems like a very odd and hard to nail down issue, sorry that this is dragging out so long.

[swiggu]

Nope. At this point, I don't even get the occasional GPU driver crash and recover, it's just purely hard hangs/freezes. Other source entries include a lot of "SideBySide" errors, which appear to be DropBox related, "NvStreamSvc", "SNMP", "WMI" and "WLAN-AutoConfig". The error displayed around the time of a freeze is always "Kernel-Power".

Yeah, it's quite odd. And yeah, it's dragging out, but hopefully we'll get there, and I'm very grateful for the help I'm being given.

[Calum]

Mhm, the kernel power event log entry is "normal" in that whenever the machine unexpectedly powers off or restarts, that will be entered.  The others seem pretty unrelated to this issue too.

What exact kit of RAM do you have, do you have a link to it on the manufacturer's site at all?
It might be worth manually setting the speeds, voltage and timing, just in case, although I'd expect a marginal issue with the RAM to manifest itself outside of the problems you're currently having.  Passing Memtest is a good start but it doesn't always guarantee 100% stability, more like 99.9% so still a bit of a long shot but perhaps worth a try.  Past that, we could try tweaking a few other voltages, but again I have my doubts.

Sorry to sound pessimistic, I'm still hopeful we'll get this resolved, just not quite sure how at this exact moment.

[swiggu]

Two of these I believe: http://www.kingston.com/datasheets/KHX1600C9D3B1K2_8GX.pdf.

I ran Memtest for a couple of passes but it came out without any problems.

It's alright. I've browsed other forums for similar problems, but the different causes and solutions are literally within the hundreds, so it's hard to even find a proper place to start looking. I've been going at it systematically, but at this rate I could end up buying new parts for twice the amount of money if I can't find the root of the issue.

[Calum]

The Kingston RAM is usually solid stuff, I've used hundreds of the "new" (2010-plus) HyperX kits and was so impressed I made it my go-to kit (well, either the blue/grey/red/black as they're all the same) - I don't know that I ever ended up with a failed kit.  We did however have some odd compatibility issues with one or two MSI boards, which is why I'm now thinking a BIOS update might be the next thing to try.  It looks like there have been a few fixes related to system stability in the releases after the initial version, so go ahead and update to version 10.3 (or A3)  if you're not already on that.
If you're already on that BIOS, or the freezes continue, try setting your RAM timings and speeds manually.  It looks like you can do this under the OC menu in the BIOS, set the DRAM Frequency to 1600MHz, then set DRAM Timing Mode to Link and enter Advanced DRAM Configuration.
You'll want to set the timings to 9-9-9-28-2, in order that's as follows: CAS Latency, RAS to CAS, RAS Precharge, tRAS, and command rate.  It could be worth setting tRC to 37 while you're there.
Set the DRAM voltage to 1.65V, and disable spread spectrum.

See how you go on with that, hopefully some or all of that will at least help.  If you run into any issues, let us know, I'm working off a downloaded manaul for your board so some settings may not be exact if anything has changed since the manual was published.

[swiggu]

Yeah, I bought that RAM based on really positive reviews (pretty much every other part as well).

I've updated my BIOS to the latest version already, so that's that. It could indeed be compatibility with my MSI board (read some stuff about that).

I'm gonna try to configure my BIOS the way you suggested I do, thanks a lot. I'll tell you more when I'm done.

[swiggu]

Changed 'em all except for tCAS, which I couldn't find. On a side note, do I have to select OC Genie mode for the changes to take effect?

[Calum]

Is there a value set to around 28 (probably 24 or 27)?  If so, that may be tRAS, some BIOSes name them differently.
You shouldn't need to select OC Genie mode from what I can see in the manual, looks like just setting DRAM Timing Mode to Link instead of Auto, and setting the voltage manually, should be fine.  From what I remember of MSI boards, OC Genie is their automatic overclocking, which I would avoid.

[swiggu]

So far so good after making those changes. Also switched to Chrome as my browser because I felt Firefox was really slow and potentially a cause of the freezes. Once again I'll post in around a week if it's entirely gone.

[Calum]

Keeping my fingers crossed for you

[swiggu]

Happened again just now unfortunately. :/

Alternate offers full testing of a machine if I bring the entire thing over to them so I might try that out, unless there's any other things I could try out myself.

[Calum]

Ahh, gutted I was really hoping this was it this time.
I think taking the machine in to Alternate if that's something they offer is a sound idea at this point, not really sure what else to suggest.  Sometimes the only way is to have the machine somewhere with a ton of parts to swap out.

[swiggu]

I'll be doing that soon then. I'll tell you what the problem was when I have. Thanks for all the help nonetheless! :^)

[swiggu]

Alright, so after two months of having Alternate look at my PC they found nothing, even though when I brought it back home it started acting up soon after with the exact same issue: hard freezes I can't recover from. I did some more researching and it turns out my motherboard might be the culprit. Apparently it's a really *censored* motherboard that has bad support for my range of CPUs and its "C-Caps" or whatever those may be will overheat extremely fast unless cooled, and I don't have any cooling on my mobo aside from on the CPU itself (the fan it came with). Seeing how it's always kind of a "snowballing" effect of freezes, where it may happen after four hours and then in half a minute or even less, it may be heat-related, and it's an area I can't monitor for temperatures using software. Should I consider a new motherboard?

[swiggu]

bump.

[patio]

MSI says it all...

[swiggu]

Are they considered bad? I honestly had no idea; I went and bought it based on user reviews which seemed fairly positive.

Would you suggest I purchase a new motherboard and, if so, what would you recommend?

[patio]

Not quite sure everything else has been ruled out...without spending any more on diagnosing however perhaps you could borrow some equivalent RAM stiks to swap in there for a day or so to completely eliminate  the RAM as the culprit...

[swiggu]

Alright, so I tested with some alternative RAM which didn't solve anything. I checked back into the event manager and kept noticing there's a WLAN-related error, so I decided to remove my network card and try using a LAN cable and a Netgear wireless USB network adapter and the freeze hasn't occured for over a week now, so I'm guessing that was the problem here. Is there anything I could do to fix the malfunction adapter or is it just faulty period?

[patio]

I know of no fixes for wireless adapters...
All hardware fails at some point.

[swiggu]

Guess I could try buying another PCI-e net adapter, but it could be that the PCI-e is faulty and that causes it as well.

[patio]

If the new one does the same return it...then start shopping for a new MBoard.

[swiggu]

Yeah, this motherboard is rather problematic in many other aspects it seems. For example, the USB network adapter I'm using right now doesn't function well on the motherboard's USB ports and I get high ping and stuff but works perfectly on the front panel USB. Only the mouse and keyboard seem to work well on the motherboard's USB ports. :/