|
Problem description: Recently my computer has been sort-of crashing. The displays have been going black. It seems to be happening more frequently. Three times this evening, the displays went black, accompanied by a nasty buzz in my headphones. This is while being connected to Discord and playing a game with some folks. The first time, I couldn't hear anything any more from Discord, but pressing my "toggle mute" button made the characteristic beeps. The second time, I was still fully connected and could hear everyone and tell them "yup it happened again!" So the computer hasn't crashed, it's just the display. I also noticed lots of fan noise when it happened. I DL'd a temperature logger. Nothing unusual. I've just totally uninstalled and reinstalled the display drivers. That was between crash #2 and crash #3. And that's after updating the driver last night. WTF? Dying graphics card? It's only three years old, it shouldn't be dying this quickly  edit: I think maybe it has something to do with how much load I'm putting it under? I played about half an hour of Hardspace Shipbreaker with no problems, then started playing Planetside 2 and it crashed within an hour (crash #1). The next two crashes were after trying to start up PS2 again. I've been browsing the forums since then, with no issues. Note that I've also played Planetside 2 just fine for months and months on this machine. Recent changes: None. -- Operating system: Windows 10 Home v1909. System specs: CPU: AMD Ryzen 5 3600 GPU: Zotac nVidia GeForce 1070 PSU: Corsair CX 650. Motherboard: MSI B450-A PRO MAX RAM: 16gb Storage: Samsung 500GB EVO Plus NVME, two Western Digital magnetic drives. Location: Australia I have Googled and read the FAQ: Yup. Hyperlynx fucked around with this message at 12:31 on Aug 5, 2020 |
#
?
Aug 5, 2020 12:00
|
|
|
|
| # ? Apr 23, 2024 21:15 |
|
|
Try resetting your BIOS to defaults, including removing any DOCP RAM overclocking you may have. If the error goes away, it's probably memory or memory configuration. If it does not, it's probably GPU. A good stress test for the GPU would be FurMark. https://geeks3d.com/furmark/
|
|
#
?
Aug 5, 2020 19:40
|
|
|
DOOMocrat posted:Try resetting your BIOS to defaults, including removing any DOCP RAM overclocking you may have. If the error goes away, it's probably memory or memory configuration. If it does not, it's probably GPU. I found evidence in the logs that it was the display driver crashing. After uninstalling the display drivers with a special tool, then reinstalling, it hasn't crashed since but who knows. I tried FurMark just now. It didn't crash, but the GPU temp reached 80 degrees C, which sounds kind of hot? I don't know.
|
|
#
?
Aug 8, 2020 12:04
|
|
|
Hyperlynx posted:...but the GPU temp reached 80 degrees C, which sounds kind of hot? I don't know. That's fine for a GPU stress test.
|
|
#
?
Aug 9, 2020 22:23
|
|
|
Well gently caress. Wednesday evening I set a bootable USB memory test running. 18 hours later I checked the results and the RAM is fine. I do some other stuff on my computer, everything is fine. This evening (Friday), I switched on my computer and the screens didn't wake up. Fans went, CD drive made CD drive noises, but no computer. And the lights on my keyboard and my mouse didn't light up. gently caress. gently caress gently caress loving gently caress. So, what, now the mobo's dead? I had something similar happen around the start of last year too. I replaced the PSU, the mobo, mobo died again, so I got an electrician in who said all my power points are fine. Then I replaced the mobo, the CPU, and the case for good measure. I'm pretty sure I posted about it here, but can't find the thread. What the gently caress. I've only just replaced, well, most of my machine. I didn't replace the graphics card, but it was only 3 years old and I understood Zotac to be a reliable brand. And Corsair have a good reputation. What the hell is going on? I mean, at this point, wtf am I meant to do but bury the entire thing, conduct an exorcism, put up Shinto seals on the grave to stop the evil spirits escaping, and then build a new computer entirely from scratch? With a UPS for good measure? 
|
|
#
?
Aug 14, 2020 13:53
|
|
|
Hyperlynx posted:So, what, now the mobo's dead? I had something similar happen around the start of last year too. I replaced the PSU, the mobo, mobo died again, so I got an electrician in who said all my power points are fine. Then I replaced the mobo, the CPU, and the case for good measure. I'm pretty sure I posted about it here, but can't find the thread. I'd try to see if the onboard GPU works but there's a decent chance your PSU died or isn't putting out enough power for the GPU any longer. The Corsair CX line is not the best. PS It wouldn't be hard to believe a three year old Zotac card died either. I've seen it before. Zogo fucked around with this message at 21:57 on Aug 14, 2020 |
|
#
?
Aug 14, 2020 21:52
|
|
|
So.... turns out my computer problems might have been a simple loose cable all along. I opened up the case to check the diagnostic LEDs (after a kind stranger elsewhere reminded me they exist) and noticed the main motherboard power cable didn't look like it was seated right. I powered off, pushed it in harder, powered on and voila, here we are  Phew! My system isn't totally hosed. Hopefully that solves my other problem, too!... But, for future reference, what parts would you consider most reliable?
|
|
#
?
Aug 15, 2020 04:10
|
|
|
|
| # ? Apr 23, 2024 21:15 |
|
|
Hyperlynx posted:But, for future reference, what parts would you consider most reliable? In recent years I've seen the most GPU failures from EVGA cards but they seem to have a good reputation so it could just be an anecdotal/sample size issue. Gigabyte was doing unethical things regarding voltage readings on motherboards a few years back. So that hurt their reputation. As far as PSUs this is from the PC building thread: "Power supplies have roughly doubled in price due to the current trade war. Cheap PSUs are now pricey and good ones are very pricey. Its recommended to get an 80+ Gold (or better) PSU with a 7 to 10 year warranty. While not an absolute sign of quality, long warranties are indicative of the manufacturers confidence in their product. Currently this includes the Seasonic Focus and Focus Plus Gold lines, Corsair TMx, RMx 2018 and RMx 2019 lines, and EVGA Supernova G1/2/3 lines." Things can change from year to year so it's good to just stick to highly rated products and ask in those hardware megathreads etc.
|
|
#
?
Aug 15, 2020 05:10
|
|