- May 19, 2011
- 18,982
- 12,105
- 136
Spec:
AMD Phenom II X4 960T, running completely at stock
4GB DDR3
ASUS M4A89GTD PRO/USB3, latest BIOS
GeForce 760
Samsung 850 PRO 128GB (boot)
Corsair VX550W
Windows 10 (up-to-date)
The history is a bit convoluted. The graphics card and PSU come from my previous 'spare PC' which was 100% stable (Nehalem build) and I gave to my brother. I kept those bits because he didn't need them and I do need a PC with a bit of graphical oomph and accelerated H264 for DVD ripping. The rest of the hardware was from a computer build I did for a customer back in 2011 and it had a couple of stability incidents (one I mentioned here, the PC had been crashing but the CPU HSF was a bit dusty, I also deactivated the two additional cores that had been activated 'just in case'). After the first, the PC was fine for a few months. The second involved graphical glitching and crashing (onboard graphics) and random BSODs.
I did a clean install of Windows 10 onto a spare SSD I had and started using this computer for ripping DVDs/BRs with Handbrake. Sometimes GPU-assisted, sometimes not. I got the occasional BSOD (IRQL_NOT_LESS_OR_EQUAL, SYSTEM_THREAD_EXCEPTION_NOT_HANDLED, 0x00000119, BlueScreenViewer cites various DLLs as the cause, mostly system ones). I tried switching graphics card drivers too. The PC has handled ~20 hr long software encodes, and it's not as if it only fails with GPU workloads.
Today however has been interesting: Four BSODs while idle (and I haven't given it any work to do), including KERNEL_SECURITY_CHECK_FAILURE.
I've run memtest86+ 5.01 overnight at least once while I've suspected stability issues, no errors. I've Prime95'd it on various settings for at least half an hour, but now I think we can safely say it's not load related.
Previously I've wondered if the extra two cores were an issue but since I've had quite a few crashes since setting it back to a standard 960T, either the CPU is faulty or I can rule it out.
There's nothing particularly of interest in the event log around the time of the BSODs or otherwise.
I'm inclined to believe that the board has had it, but if anyone has any ideas here for testing/narrowing down the problem, let me know. I hate sending OK hardware to the dump!
I'm going to run memtest 4.3.7 (the machine isn't UEFI so that's the latest version) just in case the memory has become 'more' faulty recently.
AMD Phenom II X4 960T, running completely at stock
4GB DDR3
ASUS M4A89GTD PRO/USB3, latest BIOS
GeForce 760
Samsung 850 PRO 128GB (boot)
Corsair VX550W
Windows 10 (up-to-date)
The history is a bit convoluted. The graphics card and PSU come from my previous 'spare PC' which was 100% stable (Nehalem build) and I gave to my brother. I kept those bits because he didn't need them and I do need a PC with a bit of graphical oomph and accelerated H264 for DVD ripping. The rest of the hardware was from a computer build I did for a customer back in 2011 and it had a couple of stability incidents (one I mentioned here, the PC had been crashing but the CPU HSF was a bit dusty, I also deactivated the two additional cores that had been activated 'just in case'). After the first, the PC was fine for a few months. The second involved graphical glitching and crashing (onboard graphics) and random BSODs.
I did a clean install of Windows 10 onto a spare SSD I had and started using this computer for ripping DVDs/BRs with Handbrake. Sometimes GPU-assisted, sometimes not. I got the occasional BSOD (IRQL_NOT_LESS_OR_EQUAL, SYSTEM_THREAD_EXCEPTION_NOT_HANDLED, 0x00000119, BlueScreenViewer cites various DLLs as the cause, mostly system ones). I tried switching graphics card drivers too. The PC has handled ~20 hr long software encodes, and it's not as if it only fails with GPU workloads.
Today however has been interesting: Four BSODs while idle (and I haven't given it any work to do), including KERNEL_SECURITY_CHECK_FAILURE.
I've run memtest86+ 5.01 overnight at least once while I've suspected stability issues, no errors. I've Prime95'd it on various settings for at least half an hour, but now I think we can safely say it's not load related.
Previously I've wondered if the extra two cores were an issue but since I've had quite a few crashes since setting it back to a standard 960T, either the CPU is faulty or I can rule it out.
There's nothing particularly of interest in the event log around the time of the BSODs or otherwise.
I'm inclined to believe that the board has had it, but if anyone has any ideas here for testing/narrowing down the problem, let me know. I hate sending OK hardware to the dump!
I'm going to run memtest 4.3.7 (the machine isn't UEFI so that's the latest version) just in case the memory has become 'more' faulty recently.