Hi, long time lurker here but this is my first post. :whistling:
I built this system in July of last year and it was running perfectly up until two months or so ago, when it started crashing after a random amount of time whenever I played any games with 3D graphics. It's only seemed to have gotten worse since then and I've tried so many things to try to fix the crashes but each thing only seemed to 'work' for a week before the crashes resumed and then it never worked a second time. I'm at my wit's end at this point but I think I might have pinpointed this whole ordeal to the motherboard being a cockup. Unfortunately I don't have any experience in fubar'd motherboards so I'd like a second opinion.
System Specs
hastebin link.
CPU never goes over 80 degrees centigrade unless I really stress on purpose it (e.g. with Prime95's torture test). Idles at 32 or so degrees.
GPU has only gone above 60 degrees a few times while under load. Once was when it hit 70 while benchmarking and every other time it only went about 2 or 3 degrees above 60.
What's Wrong
Whoo boy. This all started with the crashes. The sound would freak out for a couple of seconds as the screen went black and then the computer would power off and back on again. Windows told me that it was bluescreening and WhoCrashed told me that the BSODs were always stop code 116, VIDEO_TDR_ERROR and that the crash was 'likely' caused by nvlddkm.sys. I made sure that the drivers were up to date (they were, and had not been updated recently) and reseated the GPU. With this, the problem seemed to go away for a few days.
When it came back, it came back with a vengeance though. The crashes went from randomly during a game to one or two seconds after a game started up every time a game started up. Cleanly installing the drivers (I tried both Nvidia's clean install option and using device driver sweeper) did nothing.
I RMA'd the graphics card and while I was running on only my integrated graphics the crashes went away completely. I also ran memtest and all four passes came back with no errors.
When I got the new card back (ASUS replaced it) the issue was still there, except after the first crash the machine would power on but wouldn't boot. No post, no bios splash, nothing. Went through the troubleshooting motions and it turned out reseating the RAM seemed to get the PC to boot, but still crash when launching games with 3D graphics. I experimented with the RAM a little bit and it turned out that by putting one particular stick in the first slot the computer would go into an 'I have no RAM' loop, and putting it in the second slot with another stick in the first would lead to the crashes. I ran the computer for a week with just the 'good' stick of ram and didn't run into any crashes. Maybe memtest had just been wrong somehow?
Buuut then they came back again, with or without the 'bad' stick. Ran another four memtest passes and still no errors.
I installed Steam on the Linux drive that I previously only used for dev stuff, thinking that maybe I'll be lucky and it's only Windows being Windows. Played a few games with no crashes and got pretty excited. Went back to playing them and Windows to confirm and... no crashes.
A few days later (today) I once again tried to launch something on Windows only to have my screen freeze and artifact up. Manually turned off the computer and tried to boot into Windows again, only to have everything turned to artifacts after the Windows splash screen, followed by a black screen.
I tried booting into the Linux drive and the same thing happened on that OS.
Booting into BIOS did not result in a black screen but everything was still covered in artifacts.
I took out the GPU again and went back to my integrated graphics and everything looks fine on both operating systems and BIOS.
I'm now pretty sure that the next reasonable option to look at would be the motherboard being FUBAR, but broken mobos are one of the things that I have zero experience in, so I'd like that second opinion from people who hopefully know more. I also have no experience with borked PSUs, so it could be that too I guess.
Oh, and if there's anything else I might be missing please tell me. Considering how everything seems to work completely fine when there's nothing in the PCI-e slot, I don't think it's the CPU. I've been monitoring temps pretty religiously for a while and there's nothing to lead me to suspect that the computer is overheating, especially since most of the crashes happen before the computer even gets under load.
I'd think that ASUS managed to send me two shite GPUs in a row but that wouldn't explain the RAM thinking it's broken for a week and then changing its mind...
I've never had a hardware problem before where the problem was so damn hard to reproduce.
tl;dr - Two operating systems, two GPU's and eight memtest passes later, I can only suspect that my computer may be haunted.
Maybe I should just start RMAing parts like mad until I get a rig that works.
I built this system in July of last year and it was running perfectly up until two months or so ago, when it started crashing after a random amount of time whenever I played any games with 3D graphics. It's only seemed to have gotten worse since then and I've tried so many things to try to fix the crashes but each thing only seemed to 'work' for a week before the crashes resumed and then it never worked a second time. I'm at my wit's end at this point but I think I might have pinpointed this whole ordeal to the motherboard being a cockup. Unfortunately I don't have any experience in fubar'd motherboards so I'd like a second opinion.
System Specs
hastebin link.
CPU never goes over 80 degrees centigrade unless I really stress on purpose it (e.g. with Prime95's torture test). Idles at 32 or so degrees.
GPU has only gone above 60 degrees a few times while under load. Once was when it hit 70 while benchmarking and every other time it only went about 2 or 3 degrees above 60.
What's Wrong
Whoo boy. This all started with the crashes. The sound would freak out for a couple of seconds as the screen went black and then the computer would power off and back on again. Windows told me that it was bluescreening and WhoCrashed told me that the BSODs were always stop code 116, VIDEO_TDR_ERROR and that the crash was 'likely' caused by nvlddkm.sys. I made sure that the drivers were up to date (they were, and had not been updated recently) and reseated the GPU. With this, the problem seemed to go away for a few days.
When it came back, it came back with a vengeance though. The crashes went from randomly during a game to one or two seconds after a game started up every time a game started up. Cleanly installing the drivers (I tried both Nvidia's clean install option and using device driver sweeper) did nothing.
I RMA'd the graphics card and while I was running on only my integrated graphics the crashes went away completely. I also ran memtest and all four passes came back with no errors.
When I got the new card back (ASUS replaced it) the issue was still there, except after the first crash the machine would power on but wouldn't boot. No post, no bios splash, nothing. Went through the troubleshooting motions and it turned out reseating the RAM seemed to get the PC to boot, but still crash when launching games with 3D graphics. I experimented with the RAM a little bit and it turned out that by putting one particular stick in the first slot the computer would go into an 'I have no RAM' loop, and putting it in the second slot with another stick in the first would lead to the crashes. I ran the computer for a week with just the 'good' stick of ram and didn't run into any crashes. Maybe memtest had just been wrong somehow?
Buuut then they came back again, with or without the 'bad' stick. Ran another four memtest passes and still no errors.
I installed Steam on the Linux drive that I previously only used for dev stuff, thinking that maybe I'll be lucky and it's only Windows being Windows. Played a few games with no crashes and got pretty excited. Went back to playing them and Windows to confirm and... no crashes.
A few days later (today) I once again tried to launch something on Windows only to have my screen freeze and artifact up. Manually turned off the computer and tried to boot into Windows again, only to have everything turned to artifacts after the Windows splash screen, followed by a black screen.
I tried booting into the Linux drive and the same thing happened on that OS.
Booting into BIOS did not result in a black screen but everything was still covered in artifacts.
I took out the GPU again and went back to my integrated graphics and everything looks fine on both operating systems and BIOS.
I'm now pretty sure that the next reasonable option to look at would be the motherboard being FUBAR, but broken mobos are one of the things that I have zero experience in, so I'd like that second opinion from people who hopefully know more. I also have no experience with borked PSUs, so it could be that too I guess.
Oh, and if there's anything else I might be missing please tell me. Considering how everything seems to work completely fine when there's nothing in the PCI-e slot, I don't think it's the CPU. I've been monitoring temps pretty religiously for a while and there's nothing to lead me to suspect that the computer is overheating, especially since most of the crashes happen before the computer even gets under load.
I'd think that ASUS managed to send me two shite GPUs in a row but that wouldn't explain the RAM thinking it's broken for a week and then changing its mind...
I've never had a hardware problem before where the problem was so damn hard to reproduce.
tl;dr - Two operating systems, two GPU's and eight memtest passes later, I can only suspect that my computer may be haunted.
Maybe I should just start RMAing parts like mad until I get a rig that works.