Random crashes, shut downs... Area-51 ALX

Hello morblore,

Thank you for your answer.

I already did the complete scan from the page you linked (it's the same as the AlienAutopsy tool), and I just did the Blue Screen Analyzer.

Here are the results :

The test started and only took a few seconds to complete (even if it's written it should take about 20mins).

I'm going to try taking out the GPU one at the time, I'm just no so sure about how to do it so that I don't destroy everything : where do I put the GPU that I get out of there ? I guess I can't just put it on a corner of my desk, that thing is really sensitive...

Hi everyone,   I'm experiencing random shut downs while playing video games on my Area-51 ALX.   It always starts the same way : the screen starts to flash (gets black for a few seconds, I can still hear the sounds of the game, then the screen gets back on). The computer "freezes" when this happen (I tried typing something while the screen was black, when it came back nothing happened), so I know the problem doesn't come from my Monitor. When this happens, I sometimes get a little Windows pop-up : "NVIDIA Graphics Driver version xxx.xx crashed and was recovered." or a messages telling me something about a Windows video response delay.   This can happen many times in a row, and take a while. Then, the screen gets black again, but I can hear that the sound is bugged, and the screen doesn't turn on again (instead it writes : "VGA - No Signal"), the computer makes a lot of noise (he sets all the fans on max for a few seconds), then he turns off and restarts.Alternatively, it also happened that I just get a blue screen.   I then get the usual "Do you want to boot in safe mode" message. Sometimes, he starts crashing again right after that, sometimes, it can take another hour before crashing again.   It's really strange because sometimes I can play BF3 for hours in 3D on max settings without crashing, and later on crash while I'm using word.   At first, I thought it was a heat problem, as hinted here : http://www.dell.com/support/troubleshooting/us/en/19/KCS/KcsArticles/ArticleView?c=us&docid=429834&doclang=en&l=en&s=bsd&cs= (I do have Dual GTX580 in SLI on slots 1 and 3).   But : 1.) The problem already happened in December, when it was really cold in my room (after that, I just couldn't get the video to work w/ a DVI cable and had to change to a VGA).   2.) I "forced" the NVIDA fans and all the other fans on 100% and tried launching a game, still didn't work.   3.) I started a game, opened HWmonitor, and noticed that all the temps (CPU and both GPUs) were around 50°C. Then the computer crashed.     I ran "PC-Doctor", he said there's no problem with my computer (even if he tells me he crashed 40 times in a day). I updated all the drivers, it still crashes. Got back to a "stable" driver version, one that worked for sure month ago : it still crashes. I cleaned the inside of my case, there was no dust at all.     I don't know what to do anymore. Has anyone experienced the same problem ? How did you fix it ?   Here's my config :   -CPU : i7 960 @3.20Ghz   -GPU : Dual GTX580 SLI   -RAM : 6Go   -Windows 7 Pro     PS : Sorry for my bad english, I'm not a native speaker.

Hello Lt_Rush3r


Sorry about your crashing. Did you run the Blue Screen analyzer from Dell? If not run it first please so we can see what is going on. Go here: http://www.dell.com/support/Diagnostics/us/en/19 


If you really think it's a GPU issue you have a few things you could try. Since you are running SLI I would try pulling one card and running only a single 580 in the top slot. Try that for a day or two and then do the same with the 2nd card. That can rule out if it's a bad card. If it only crashes with one card and not the other then you have found the issue

Hello morblore,   Thank you for your answer.   I already did the complete scan from the page you linked (it's the same as the AlienAutopsy tool), and I just did the Blue Screen Analyzer.   Here are the results :     The test started and only took a few seconds to complete (even if it's written it should take about 20mins).     I'm going to try taking out the GPU one at the time, I'm just no so sure about how to do it so that I don't destroy everything : where do I put the GPU that I get out of there ? I guess I can't just put it on a corner of my desk, that thing is really sensitive...

Ok, go here instead and download/run this one. I don't know what's up with Dells but I've been using this one for years and I know it works. There should be crash files on your system from the bluescreens, this will read them. Post the info when you get it.

BlueScreenView v1.52
http://www.nirsoft.net/utils/blue_screen_view.html



To remove the GPU. Power down the system, unplug it from the outlet and hold the power button down for 10sec. Open the side of the case, open the GPU shroud. Touch a part of the metal frame on the inside of the case. Remove both power connectors on the bottom GTX 580. Then remove the screws on the far left holding it in. Remove the SLI bridge. Then just pull out but try pulling it out very easily. If it does not move then you need to push the PCI_E tab on the back to release it. It is on the right side of the slot that the GPU is sticking in.

That's it.

Store it away from anything that can give a static discharge like carpet. DO NOT HANDLE IT ANYMORE then you have too. Side of desk is fine. I do that a lot myself at home and almost everyday at work. Never had an issue.

I'm getting this when running BlueScreenView, not sure about how to interpret it :

Thanks for the tips about the GPU. I'm going to try that tomorrow and tell you if it worked.

I'm getting this when running BlueScreenView, not sure about how to interpret it :     Thanks for the tips about the GPU. I'm going to try that tomorrow and tell you if it worked.

Hi again  

Try doing a complete uninstall of the Nvidia drivers. Go to the control panel, device manager and then the display adapters and right click the video card. Then click the uninstall and check the box to delete the driver. Once done reboot and let windows do it's thing. Once that is all done go to Nvidia and get the latest driver from them and install it.

After that go here and download and install DirectX end user: http://www.microsoft.com/en-us/download/details.aspx?id=35 


Wow you were not kidding, you have been dealing with this crash forever. Good news is I think this will fix it. Looks like it's just software. Cross fingers  One other thing. With so many crashes your Windows could be messed up. It may be a good time to start over with a fresh install but don't do that yet. Lets see what this does first. For now run the single GPU so we can rule that out. I also want to confirm you are connecting the power cables correctly.

For the cables what you want to do is to make sure one 8-pin connector on each GPU is powered by it's own 8-pin power cable. Make sure that the power cable is powering ONLY one 8-pin on ONLY one GPU. So you will have one cable powering one 8-pin and another cable powering the other 8-pin. The 6-pins can be powered by a single cable or by 2 different cables, doesn't matter just don't take power off the cables supplying the 8-pins.

I unistalled the Nvidia drivers But when I tried to install the DirectX thing, the program told me it was already installed on my computer.   I'm going to try to launch a game to see if it keeps crashing before I get a card out of there.   Should I check the power cables before I take the card out of the computer or after (is it possible that the crashes happened because the power cables were not connected the right way ? I mean, I never changed anything in this part of my computer) ?

That means the DirectX you have is up to date which is good.  Yeah, you can confirm the power cables before testing the cards. It could have been sent wrong, you never know. From your post though it does look like a driver issue.

From all the crashing it could have corrupted Windows too. It's a possibility. At this point if it was my system I would do a quick look to confirm the power cables. Then the full uninstall of the Nvidia driver like you did. Then confirm DirectX was up to date also like you did. If none of that did it for me I would reinstall Windows without a second thought. Splitting the cards up is still a good idea because that can rule out a bad GPU. The BSOD has been going on for a long time. In that time I would assume driver updates have been going on. So lets rule out a bad GPU or connection too.


Something else you can always do is open the RUN window and type DXDIAG. On the next window you will see all the system info. As you scroll thorough the windows it will tell you down below in the notes window if it sees any issue. It should say no problems found for each tab.  

I unistalled the Nvidia drivers But when I tried to install the DirectX thing, the program told me it was already installed on my computer.

I'm going to try to launch a game to see if it keeps crashing before I get a card out of there.

Should I check the power cables before I take the card out of the computer or after (is it possible that the crashes happened because the power cables were not connected the right way ? I mean, I never changed anything in this part of my computer) ?

I agree, it seems like a driver issue... But the driver crashed again even after my fresh reinstall.

Here's a screenshot of the error message :

Which means in french : "The display driver was not answering but was recovered. The NVIDA Windows Kernel Mode Driver Version 320.49 was not answering."

I had 2 crashes like this one today, but no blue screen.

Here's a screen of the windows maintenance center :

Translation : "Resolution of a problem in NVIDA Graphics Driver. The driver crashed 23 times, the last time on the 07/27/2013 at 8:10PM. A resolution procedure is available". You can see that the driver sometimes crashes without crashing the computer.

Sadly, when I click for the details and try to access the recommended solution, I get a "It seems that we can't load the website, please check that you're connected to the Internet and try again later".

You are right, I kept upgrading the drivers as soon as they came out, so they were always up to date.

I tried to run the DXDIAG, it told me there was no problem of any kind.

Let me check the power cable, and see if there is a problem.

Anyways, about the windows reinstall, can I try first booting on a different windows install ? I actually have another hard drive with another windows install on it, which I could boot on. I would like to avoid loosing data, if possible.

I agree, it seems like a driver issue... But the driver crashed again even after my fresh reinstall.   Here's a screenshot of the error message :     Which means in french : "The display driver was not answering but was recovered. The NVIDA Windows Kernel Mode Driver Version 320.49 was not answering."   I had 2 crashes like this one today, but no blue screen.   Here's a screen of the windows maintenance center :     Translation : "Resolution of a problem in NVIDA Graphics Driver. The driver crashed 23 times, the last time on the 07/27/2013 at 8:10PM. A resolution procedure is available". You can see that the driver sometimes crashes without crashing the computer.   Sadly, when I click for the details and try to access the recommended solution, I get a "It seems that we can't load the website, please check that you're connected to the Internet and try again later".   You are right, I kept upgrading the drivers as soon as they came out, so they were always up to date.   I tried to run the DXDIAG, it told me there was no problem of any kind.   Let me check the power cable, and see if there is a problem.     Anyways, about the windows reinstall, can I try first booting on a different windows install ? I actually have another hard drive with another windows install on it, which I could boot on. I would like to avoid loosing data, if possible.

It would need to have the proper drivers like the chipset but you could try it. I have personally had BSOD's that were fixed with a fresh OS install so it is something you may need to try at some point.

Was it on both cards this time when it crashed? If not pull a card and start testing each one.

3 big things need to happen here. Need to test each card by itself to confirm they are both good. Need to try a fresh OS install to rule out a windows issue which would also take care of a driver issue. Need to confirm the power connectors, that was a known issue for the system. Maybe use that other drive to back up your data?

If it still does it after all that then you are looking at a hardware issue other then the GPU's like the mobo or something.

The power connectors were a known issue ?

I opened it, checked the wires : I think they're plugged in the right way...

The 8 pin connecter is the one on the left, right ? I wasn't sure because it seems it's a 6+2. But the black cables from the 2 pin connector were going to the connector on the left (first picture above) so I guess this is the 8 Pin one. Anyways, each card has its own 6 pin and 8 pin wire.

I can also confirm there's no heat problem, there's no dust in there and the cards were cool even though the PC just crashed after a series of black screens.

I'm a bit afraid of taking one of the card out of there. These power cables are making knots, and I can't access the PCI_E tab on the second card because both cards are too close. By the way, can you tell me which card is the 1st and which one is the second one on the control pannel ? I think the 2nd one is the one below the other (because the video cable is connected to the upper one), but I noticed something which could be meaningful...

I did put the graphics cards' fans on max speed via the NVIDIA control pannel to be sure it's not a cooling problem, remember ? Well... I noticed that every single time the "Nvidia driver mode kernel version xxx.xx" message appears (and tells me that the driver rebooted) or that the screen gets black for a second, the card #1 gets back to the default setting (where the fan speed is controlled automatically to follow the heat increase), but not the second card, which stays on max speed fan setting. That's kinda odd.... I could try to get that "Card #1" out of there first, but I don't know which one it is.

Also, is there any reason why in the "Driver & Downloads" tab in the support area of the dell website http://www.dell.com/support/troubleshooting/us/en/19/servicetag/92Z265J (there's my service tag in the link, I don't know if I should copy it here on a forum), the only driver in the "video" section is this one ?

While I was searching for answers concerning my bug before I asked here, I've read on a forum that some companies use slightly "modified" cards for their computers, and that maybe the bug could be because I'm running NVIDIA drivers on an Alienware-modified card. So I'm asking myself why the latest driver on this website is 1 year and a half old. Is it because it's the only one which is stable ? Also, I wanted to try to install it but I'm not sure about a detail : does the A03 in the driver's name stand for the BIOS number ? Because I updated my BIOS something like a year ago to the A10 version...

I'm backing up all of my data right now to the second hard drive, I'm going to do a fresh install of windows later today or tomorrow.

Test each card by itself in the top slot. Use both yellow and blue cables. Clean install latest (non-beta) driver.

Always use as many different colored power supply wires as possible (dig around in case to find them all).

In the past, Dell/AW OEM cards used a standard "Reference Design" so you can use drivers from nVidia.com

I'm backing up all of my data right now to the second hard drive, I'm going to do a fresh install of windows later today or tomorrow.

When you clean install, install x58 chip drivers, but try NOT installing Intel-RST (leave it off machine completely).

 http://en.community.dell.com/owners-club/alienware/f/3746/p/19376654/19866250.aspx#19866250

At this point you really need to try each card separately and even a fresh OS install after that.  Please try using a single card for now..

You may need to pull the top card first, then bottom and then put the top card back in. There are many videos on you youtube showing how to pull a gpu. If you think you need help then watch one.

We can't really do anything else here until you do that. It needs to be tested. Once you have backed up everything you can give the OS reinstall a shot. Please post back with your results. I'm following this thread so don't worry if you takes you a couple days. I'll be here.

Good luck

Hi.

This exact same problem has happened to me today as of 3pm. Constantly happening every time I try to play any 3d game since. I have made no updates to anything on the pc since Saturday, so I am at a loss as to the cause.

I have noticed a constant error in the event log 'Kernel-Power' Event ID 41 Tastk63

Not sure if this is any help.

Bizzare as all games and system ran perfectly fine until today.

My graphics cards were replaced in March with Dual GTX 560s

system:

Win7 Ultimate,

i7 CPU 960, 3.20Ghz

RAM 6GB

250 SSD & 3TB HDD

Creative x-Fi

Hi.     This exact same problem has happened to me today as of 3pm. Constantly happening every time I try to play any 3d game since. I have made no updates to anything on the pc since Saturday, so I am at a loss as to the cause.     I have noticed a constant error in the event log 'Kernel-Power' Event ID 41 Tastk63     Not sure if this is any help.     Bizzare as all games and system ran perfectly fine until today.     My graphics cards were replaced in March with Dual GTX 560s     system:   Win7 Ultimate,   i7 CPU 960, 3.20Ghz   RAM 6GB   250 SSD & 3TB HDD     Creative x-Fi

Hi Seraphs

Read through what I posted for the OP if you haven't yet and follow the steps. First thing you need to do is is look up the BSOD code. That helps a bunch. You should also run the PSA test in the BIOS> reboot and hit f12 on BIOS screen and then scroll down to PSA. Do the full test.