My PC has been BSODing recently. It started after I restarted my PC a few days ago. I restart it maybe once a month so it could have been anything. The only change I remembered making was changing the page file usage from 2gb-8gb to system managed, which put it to 32gb. The page file was filling up even though there was tons of ram free, so i figured i'd give it some more because I didnt know why it was filling up. I changed it back and the errors kept going. I turned page file off completely because I have 32gb of ram and it still happened. This PC is running windows 10 enterprise edition and its noteworthy that it has an areca raid card in it with a couple raids (raid 10 and raid 6). It also has an AMD 7970 gfx card. The PC is used for gaming and as a plex server. The first recent BSOD happened while playing WoW. Then after the first BSOD it was chaining them immediately after logging in and starting some programs (everything.exe - comes up in some of the dumps) till I went into safe mode for a while (was backing up my PC).
I took a shot at analyzing the crash dumps, half point to memory corruption so I ran memtest overnight and it only came up with a few warnings (no solid errors). The warnings were for this case: http://www.memtest86.com/troubleshooting.htm#hammer My ram is not overclocked (nothing is), but just to make sure I underclocked my ram from 1600mhz to 1333mhz. It happened again after that. I ran a check disk and it was clean (SSD is a few months old). After taking it out of safe mode, underclocking my ram, and waiting 18 hours or so since the last BSOD, it was running WoW fine for a few hours until it crashed again.
I ran some graphics card memory tests. One of the tests (MemTestCL) came up with a ton of errors for one specific test (random), but people online doubt the program's abilities. I ran most of these here: https://www.raymond.cc/blog/having-problems-with-video-card-stress-test-its-memory/ The first program there only tests 1 gig of ram and forcing it to 3gb didnt work. It came up with no errors for 1gb of ram. The 4th (EVGA OC Scanner X) came up with 0 artifacts. One interesting thing here is that my pagefile filled up completely during one of these tests.
3 of the dumps in windbg point to: FAILURE_BUCKET_ID: MEMORY_CORRUPTION_LARGE_32
The others point to DEFAULT_BUCKET_ID: WIN8_DRIVER_FAULT --- Probably caused by : ntkrnlmp.exe ( nt! ?? ::FNODOBFM::`string'+12cb8 )
I'm still looking into this but I'm curious what you guys find as a possible root cause here. Since memtest came back fine(for the most part), I'm not sure why I'm getting memory corruption issues.
I'm currently updating my drivers and windows to see if that helps anything. I'm also going to replace the fans on my gpu since one of the doesnt spin as fast as the other. I'm tempted to just reformat and see if the issue goes away...
Moved from Windows 10 to BSOD
Edited by NickAu, 19 January 2017 - 03:42 AM.