Computer failure. Solved. Now playing with 1070 ti!

245

Comments

  • frank0314frank0314 Posts: 14,708
    edited June 2018

    People tend to underestimate the amount of fans you need to keep everything cool. I have liquid cooling but that is only for the CPU. I needed to put 3 other cooling fans and 1 exhaust fan, not including the fan that came with the case for the HD.

    Post edited by frank0314 on
  • ebergerlyebergerly Posts: 3,255
    GPUs, as well as most electronic equipment, is usually designed to be self-protecting from heat. When temps get near, say, 100c or more it will shut down.
  • frank0314frank0314 Posts: 14,708
    edited June 2018

    My computer rarely goes above 36C, even on heavy loads.

    Post edited by frank0314 on
  • Seven193Seven193 Posts: 1,141

    Of course, people who build their own custom PCs have the most opportunities to add 5 system fans or whatever.  But, if your PC isn't customizable, then you're stuck with what you got, so try to at least keep it clean of dust and cat hair. :)

  • frank0314frank0314 Posts: 14,708

    It takes nothing to add fans if you have open spots which is what can be the problem. Someone with no computer or mechanical experience can put a fan into a computer.

  • kyoto kidkyoto kid Posts: 41,851
    ...I recently added three more fans to my old Antec P-193. 7 total.
  • AnotherUserNameAnotherUserName Posts: 2,727
    wolf359 said:

    Your PSU is one of those electrical components that can literally become "ill" and die slowly, giving you some forewarning of its impending demise.
    In my recent experience with an old used HP compaq
    the Green indicator light comes on slowly instead of instantly
    like on a "healthy" PSU.
    Of course we are all  merely speculatingblush
    Is there a local PC hardware shop where you can take  it for a proper offline dignostic??

    Yah... Id like to avoid taking it to a shop though. Theres no good middle ground around here. Its either geek squad, or trendy trend trend, charge you an arm and a leg 'cause our rent is super high sort of computer shop...

  • AnotherUserNameAnotherUserName Posts: 2,727
    Dave230 said:

    Im now trying to figure out what has failed. Anyone have some pointers for what I should look for?

    Thanks.

    If the computer boots up ok again after waiting a hour or so, then the video card probably overheated.
    The only thing you can do is wait.

    Reasons for overheated card:
    - Bad airflow.  Or not enough fans.
    - Bad fan.  Check to see if all fans are spinning.

    The most important fan is the fan on your graphics card.  If it's not spinning, then the heat will kill your card permanently within 20 minutes.
     

    I'll have to check that. There are alot of fans in the computer. Its one of those alienwares that is triangular shaped to promote airflow. I think it has three fans in it in addition to the card fan.

  • AnotherUserNameAnotherUserName Posts: 2,727

    nVidia has been putting out bad drivers since the 90s.

    Ok so, I have had issues with nvidea drivers before...

    What I used to see, before all this, was a notification that a driver kernel stopped working. All of this has transpired while rendering and it always happened when the render was to much for... my cuda cores..., my vram..., it was all render related.

    ...I didnt mention that this particular failure happened once before, about a week ago, while I was feeling... ambitous... , and was trying to render a close up with high subd. The computer freaked out, just like it did with this current episode, but was able to recover. The render I was trying to cook when this happened, was not only as ambitous, but more so and was also displaying multiple if not total errors. I was trying to take the render to show the issues that I was having with the render itself. I was trying to render one of Tru Forms modern homes in totality. I new it was going to be heavy for my card but I tried anyways...

    ...hindsight... cheeky

  • AnotherUserNameAnotherUserName Posts: 2,727

    I didnt work on the pc today, I couldnt, or wouldnt, really.

    Monday, The March of Sorrow (otherwise known as mowing my lawn), followed by work on the computer. Im not going to post until then, I need to clear my head. wink

  • Ghosty12Ghosty12 Posts: 2,080
    edited June 2018

    I think you will find it is the GPU as I have had pretty much the same thing happen to me a couple of years ago, was doing some iRay rendering on a GTX 570 came back to find that the system had reset itself a small screen and artifacting all over the place.. Even during post there would be artifacts, anyway I took the card out and ran on the onboard chip and everything went back to normal.. So yeah my GTX 570 had died from a comination of iRay rendering and heat..

    The one thing is that while GeForce cards are great for gaming, they are not really designed for the heavy work load you get from renderers like iRay and Octane and so on.. This is where you have the Quadro cards come in as they are designed for this type of work, just unfortunately they cost a ton..

    Post edited by Ghosty12 on
  • I didnt work on the pc today, I couldnt, or wouldnt, really.

    Monday, The March of Sorrow (otherwise known as mowing my lawn), followed by work on the computer. Im not going to post until then, I need to clear my head. wink

    I would tend to think the video card.  I had a virtually identical catastrophic failure to what you describe about two months ago and it was the video card.

    But there are ways to help troubleshoot using the Bios error beep codes.  The Bios and computer companies build in troubleshooting aides.  First things first.

    When you try and turn on your computer, does it make multiple beeping sounds, and what is the pattern?  When you turn on a computer, it should make one short beep.  If you hear more than one, there are error beep codes built into the Bios and motherboards.  You can listen to the pattern and then look up what it means.

    Here is a more full explanation.  You said you have a Dell, so that is what I linked.  You can also Google 'computer error beep codes' and you will find multiple articles.

    http://www.dell.com/support/article/us/en/04/sln293445/understanding-beep-codes-on-a-dell-desktop-pc?lang=en

    Good luck.

     

     

  • ebergerlyebergerly Posts: 3,255
    edited June 2018

    Again, I wouldn't focus on any one thing, and especially I wouldn't overestimate the likelihood it's something like a GPU failure. Yeah, it could be, but keep in mind that GPUs are designed to throttle when temps start getting high (near what they can handle), and will actually shut down the equipment if they get too high. And the sole purpose is to protect them from damage and failure. Doesn't mean it will work 100% of the time, but it sure reduces the likelihood it will outright fail due to heat. 

    Same thing goes for CPU and other electronic stuff. 

    And don't underestimate stuff like hard drive failure. If you're loading scenes or whatever from the hard drive and it suddenly stops accessing it can cause all kinds of weird stuff to happen. I've had intermittent hard drive failures that had me stumped, cuz sometimes it works fine and other times it fails to boot. 

    And power supplies, while they are often the culprit, also have internal protection (breakers or fuses) to shut down if they get overloaded or have an internal failure. Again, doesn't mean it will always work 100%, but it certainly lowers the likelihood that your system will kinda sorta work but act wonky. And it sounds like you have at least some power since you can boot into safe mode?

    Never assume, cuz it can lead you down the wrong path. Start simple, and test each component. Cuz what might seem obvious (GPU failure due to rendering) could very well be something totally different. 

    Post edited by ebergerly on
  • ebergerlyebergerly Posts: 3,255
    edited June 2018

    FWIW, here's how I'd approach troubleshooting this. The fact you mention you couldn't boot into Windows makes me think there might be something more to this.

    1. Turn off and unplug the machine
    2. Open the case and see if there's any obvious issues, like dust clogging the GPU or other fans
    3. If you have onboard video, remove the GPU
    4. Remove any other expansion cards that aren't absolutely necessary. 
    5. Unplug any other non-essential external devices other than mouse, keyboard, and monitor (and connect that to the onboard video)
    6. Close the case and boot the machine. Take note of the beeps on startup as was mentioned.
    7. Can you boot into Windows okay? If not, it's more likely you have a basic system problem, like failing hard drive, Windows problem, hardware/motherboard problem, power supply, etc., which might take some basic system/BIOS troubleshooting.  
    8. If you can, and you have Windows 10, go to Start/Control Panel/Security and Maintenance/Maintenance/View Reliability History and see what messages/errors are there for the last few days. It might (hopefully) tell you what failed. Also check to see if you had a GPU driver update recently, or other updates. This could help to determine if there might be a bad driver
    9. You can also start checking individual components, such as chkdsk for the hard drive, test the memory, etc.
    10. If all is well at this point you can start focusing on the external/expansion devices like the GPU. Add one at a time to see if everything is still okay 

    BTW, one feature I love about W10 is the Reliability History feature I mentioned above. It lists application failures, Windows failures, other failures, updates, etc. for the last week or so. And it also gives technical details to explain in more detail. In my case I learned that my TeamViewer is shutting down incorrectly, and it lists all updates I've received lately. They recently re-designed TeamViewer and apparently I haven't yet figured out how to shut it down correctly. Anyway, I tend to check that every day on my computers to see how things are going and warn me of ongoing problems. 

    Reliability.JPG
    1262 x 551 - 123K
    Post edited by ebergerly on
  • AnotherUserNameAnotherUserName Posts: 2,727
    edited June 2018

    Okay, I give up. How does one attach a monitor to an onboard chipset?

    Being vague again... my computer is an Area 51.

    Post edited by AnotherUserName on
  • AnotherUserNameAnotherUserName Posts: 2,727
    ebergerly said:

    FWIW, here's how I'd approach troubleshooting this. The fact you mention you couldn't boot into Windows makes me think there might be something more to this.

    1. Turn off and unplug the machine
    2. Open the case and see if there's any obvious issues, like dust clogging the GPU or other fans
    3. If you have onboard video, remove the GPU
    4. Remove any other expansion cards that aren't absolutely necessary. 
    5. Unplug any other non-essential external devices other than mouse, keyboard, and monitor (and connect that to the onboard video)
    6. Close the case and boot the machine. Take note of the beeps on startup as was mentioned.
    7. Can you boot into Windows okay? If not, it's more likely you have a basic system problem, like failing hard drive, Windows problem, hardware/motherboard problem, power supply, etc., which might take some basic system/BIOS troubleshooting.  
    8. If you can, and you have Windows 10, go to Start/Control Panel/Security and Maintenance/Maintenance/View Reliability History and see what messages/errors are there for the last few days. It might (hopefully) tell you what failed. Also check to see if you had a GPU driver update recently, or other updates. This could help to determine if there might be a bad driver
    9. You can also start checking individual components, such as chkdsk for the hard drive, test the memory, etc.
    10. If all is well at this point you can start focusing on the external/expansion devices like the GPU. Add one at a time to see if everything is still okay 

    BTW, one feature I love about W10 is the Reliability History feature I mentioned above. It lists application failures, Windows failures, other failures, updates, etc. for the last week or so. And it also gives technical details to explain in more detail. In my case I learned that my TeamViewer is shutting down incorrectly, and it lists all updates I've received lately. They recently re-designed TeamViewer and apparently I haven't yet figured out how to shut it down correctly. Anyway, I tend to check that every day on my computers to see how things are going and warn me of ongoing problems. 

    So I dont know how to hook the monitor up to the onboard chipset.

    I blew out the dust, it wasnt terribly dusty... The computer makes one beep when being turned on. The alienware logo comes up as windows is loading and then I get black screen, no errors or anything. I re-boot and hit F12. I then get a message about windows not loading correctly and what do I want to do about it. I restart to go into safe mode. Safe mode is up and looks fine. I have green lines over icons and the task bar at the bottom. As I look into the computer, everything looks fine. The fans all appear to be turning and the light at the back of the psu is green although I dont know if that means anything or not. I can access my C: Drive, I can access my D: drive. I can see all of my renders that are stored there. From my untrained eye, everything looks fine inside the computer all the goofy area 51 and Nvidea lights are on.

    Thats about where im at until I can figure out how to attach my onboard chipset to my monitor... the internet has not been helpful. My dxdiag is not displaying my video card only Microsoft Basic Display Driver. No audio devices have been detected, it says because the windows audio service is not enabled.

    So im about at the same point as I was a couple days ago. I may just have to eat it and take it to a shop.

  • frank0314frank0314 Posts: 14,708

    There should be a port on the back of your computer that looks like the one you have your monitor plugged into your graphics card. This of course if you have a mother board that offers onboard.

  • Silver DolphinSilver Dolphin Posts: 1,638

    If you are rendering 3Delight then it is a CPU render. If rendering Iray it could be CPU or GPU=video card. If you are not running MSI Afterburner and have the fan running at 100% before the render and take the side off of your pc then you are overheating your Video card and killing the life of the card. It is unfortunate that Daz does not warn people that Iray is very hard on Video cards and really good cooling is a must with Iray or Octaine. Oh by the way I'm a Alienware User myself. I like the cool looking cases.

  • MescalinoMescalino Posts: 436
    edited June 2018

    If you are in safe mode that means you are using vga drivers.

    If you going into the BIOS on boot (usually ESC or F2 or F12) then navigate in just BIOS until you find where it has detected your video card. If it's not in there you video card is probably broken. 

    Would it be listed under the SATA Port devices?

    No, SATA is things like storage devices. 

    If you didn't change where your video cable is connected to after your screen artifacts when you booted to Safe Mode your video card should still work. Turn if off overnight & start it in the morning. Soon as you boot, reinstall the latest video drivers. Sometimes a bug in the driver code can cause those artifacts it doesn't have to be a broken video card.  

    How do I install from safe mode? I cant connect to the internet...

    There is a safe mode option with network support. NOTE!! If you go on the internet in safe mode any security software like virus scanners are dissabled. Be aware of that.

    Post edited by Mescalino on
  • LeatherGryphonLeatherGryphon Posts: 12,087
    edited June 2018

    If you don't already have it you might want to go to the computer manufacturer's website and see if you can find the "documentation" or "specifications" or "service manual" for your computer that has diagrams identifying all your ports and oodles of other fascinating information.  Useful for putting yourself to sleep on restless nights too.

    But if it's an old machine you might have to look in their "archive" or "obsolete" section if they have one.

    Post edited by LeatherGryphon on
  • LeatherGryphonLeatherGryphon Posts: 12,087
    edited June 2018
    Mescalino said:

    If you are in safe mode that means you are using vga drivers.

    If you going into the BIOS on boot (usually ESC or F2 or F12) then navigate in just BIOS until you find where it has detected your video card. If it's not in there you video card is probably broken. 

    Would it be listed under the SATA Port devices?

    No, SATA is things like storage devices. 

    If you didn't change where your video cable is connected to after your screen artifacts when you booted to Safe Mode your video card should still work. Turn if off overnight & start it in the morning. Soon as you boot, reinstall the latest video drivers. Sometimes a bug in the driver code can cause those artifacts it doesn't have to be a broken video card.  

    How do I install from safe mode? I cant connect to the internet...

    There is a safe mode option with network support. NOTE!! If you go on the internet in safe mode any security software like virus scanners are dissabled. Be aware of that.

    One thing that can be done in "Safe Mode with Networking" (SMwN) is to go to the official "Malwarebytes.com" website and download the free version of Malwarebytes and tell it to perform a manual scan of your system while you're still in SMwN. enlightened It's really good at identifying malware but for more sophisticated infections like viruses you really need a virus scanner like NortonSecurity, McAfee, Avast, etc.

    I've always used Norton for my antivirus software and have found that if it's already installed it doesn't run in the background while in SMwN  but you can manually tell it to run a FULL scan (which can take hours on a fully loaded system).   But that's about the only thing you can do with NortonSecurity while in SMwN..frown

    Post edited by LeatherGryphon on
  • ebergerlyebergerly Posts: 3,255
    edited June 2018

    It is unfortunate that Daz does not warn people that Iray is very hard on Video cards and really good cooling is a must with Iray or Octaine. Oh by the way I'm a Alienware User myself. I like the cool looking cases.

     

    While it might seem intuitively obvious that rendering/iray is "hard" on video cards, I'm not sure that's necessarily the case.

    Keep in mind that electronic components have a design life, and if you operate them, even continuously, within their "normal" design ratings (like temperature, frequency, etc.) they should last for their design life. A bit like how the human body can operate fine for over 100 years with a continuous temperature of 37C (98.6F). Which, BTW, isn't all that far from the temps some GPU's attain during rendering...

    So, for example, if during rendering your GPU temps get around 50C (which is probably in the "normal" or "continuous" rating), and it renders at that temp continously for a week, it should still last for its design life as long as the temps stay in the normal range. And if that design life is, say, 10 years, it's probable the GPU will be fine for 10 years. Doesn't mean it won't fail for some (other) reason, since everything has an unpredictable/probabilistic chance of failure (called "Mean Time Between Failure", MTBF).

    But whether you render for 1 minute or 1 week or 1 month at that normal temperature it doesn't (or shouldn't) really matter. The GPU's fan cooling is designed for it, and it's thermal protection is designed for it, and the components work fine at that temp.

    So yes, rendering raises the GPU's temperature, and it can be for a long time, but that doesn't mean it's a bad thing. Now, if you somehow exceed the normal range (which isn't easy if all is working as it should), and get near the GPU's "maximum" ratings, the GPU (and other electronic components) have built-in protection to either throttle performance, or shut down the computer so stuff doesn't get damaged. But the way the GPU is designed is to provide enough fan cooling so that this doesn't happen, and the temps should be continuously regulated to stay within the safe, normal, continuous range.

    So you'd need the fan/cooling to fail, and also the thermal protection to fail in order for meltdown. Unless you bought a poorly designed cheapo piece of junk I suppose.

    So if a GPU on a render machine fails it could be for many reasons, unrelated to the fact that it's used for rendering. If you want peace of mind, just monitor your temps while rendering to see if all is well. And if it fails during rendering while the temps are in the normal range, something else is probably going on.

    Post edited by ebergerly on
  • AnotherUserNameAnotherUserName Posts: 2,727
    frank0314 said:

    There should be a port on the back of your computer that looks like the one you have your monitor plugged into your graphics card. This of course if you have a mother board that offers onboard.

    Yah, I dont see one. Maybe I dont have an onboard chipset like I thought.

  • tj_1ca9500btj_1ca9500b Posts: 2,057
    edited June 2018
    frank0314 said:

    There should be a port on the back of your computer that looks like the one you have your monitor plugged into your graphics card. This of course if you have a mother board that offers onboard.

    Yah, I dont see one. Maybe I dont have an onboard chipset like I thought.

     

    I just looked up the spece for the i7-5820K, which is what you stated you have in an earlier post.

    According to the specs I read, this processor does NOT have integrated graphics.  So, assuming the stats shown by PC Parts picker are coirrect (no integrated graphics), I wouldn't be surprised if the motherboard doesn't have a VGA port.

    https://pcpartpicker.com/product/6tXfrH/intel-cpu-bx80648i75820k

    In short, yeah you might need to look at a low end (sub-$100) card to get you by for a bit, and be content doing CPU only renders, if reinstalling the drivers doesn't help/you still have corrupted graphics, and other troubleshooting doesn't point to something else.  Yeah CPU only renders are painfully slow, but at least you won't be entirely dead in the water.

    If you can snag a low end Nvidia card with Cuda cores, well at least you can do simpler renders with the card (i.e. renders that don't need a lot of video memory).

    Post edited by tj_1ca9500b on
  • AnotherUserNameAnotherUserName Posts: 2,727
    frank0314 said:

    There should be a port on the back of your computer that looks like the one you have your monitor plugged into your graphics card. This of course if you have a mother board that offers onboard.

    Yah, I dont see one. Maybe I dont have an onboard chipset like I thought.

     

    I just looked up the spece for the i7-5820K, which is what you stated you have in an earlier post.

    According to the specs I read, this processor does NOT have integrated graphics.  So, assuming the stats shown by PC Parts picker are coirrect (no integrated graphics), I wouldn't be surprised if the motherboard doesn't have a VGA port.

    https://pcpartpicker.com/product/6tXfrH/intel-cpu-bx80648i75820k

    In short, yeah you might need to look at a low end (sub-$100) card to get you by for a bit, and be content doing CPU only renders, if reinstalling the drivers doesn't help/you still have corrupted graphics, and other troubleshooting doesn't point to something else.  Yeah CPU only renders are painfully slow, but at least you won't be entirely dead in the water.

    If you can snag a low end Nvidia card with Cuda cores, well at least you can do simpler renders with the card (i.e. renders that don't need a lot of video memory).

    I actually have another card, a gtx 460. Im going to see if I can even fit that into the alienware and if so then I should be able to confirm that the gtx 770 died... I think?

  • nonesuch00nonesuch00 Posts: 18,723

    Well bummer! Sorry more money for computer parts frown

  • AnotherUserNameAnotherUserName Posts: 2,727

    Well bummer! Sorry more money for computer parts frown

    Lol. Yah. Good thing I was planning on getting a new card (assuming) within the next couple months anyhow.laugh

  • frank0314frank0314 Posts: 14,708
    edited June 2018

    Sorry for the bad news. I do custom build computers and always use the high end ASUS motherboards which don't have them either

    Post edited by frank0314 on
  • AllenArtAllenArt Posts: 7,175

    The first thing I'd do is find out what that beep code means (if it is one).

    Laurie

  • tj_1ca9500btj_1ca9500b Posts: 2,057

    Well 336 Cuda cores is better than no Cuda cores I guess.  The GTX 460 is apparently part of the Fermi family. I'm seeing a 768 MB version and a 1 GB version in this review.  1 GB will drive the monitor and Daz Studio viewport at least, but I'm guessing it won't be very helpful for most renders...

    https://www.anandtech.com/show/3809/nvidias-geforce-gtx-460-the-200-king/4

    2010 really wasn't all that long ago, except in computer hardware timeframes.  As the CPU you have is also older, while the drivers may no longer being updated, you should still be able to find some that will play nice with your system.

    As long as you have the right power connectors, I don't see any reason why the GTX 460 shouldn't fit inside your system.

Sign In or Register to comment.