I actually found this (https://github.com/nefarius/ScpToolkit/releases). Newer driver too (I believe March 27, 2016 or so?), though launching the driver itself with DS4Windows is a bad idea. Other than that, it seems stable so far, even with it running through Driver Verifier.
Seems a few different programs crashed on linux too (segfaulted). I'm gonna try to run Memtest86 again and see if I catch something this time.
Any particular reason why Memtest86+ with DDR4 memory?
memtest86+ is what we always recommend. Did some searching and found this: Memtest86+ does not appear to be actively maintained, but Memtest86 seems to have been monetized. Which would you use - Memtest86+ or Memtest86... and why? : techsupport
I don't know how accurate this is but if true then we should be recommending the free version of memtest86, not +. I'll see what I can find out.I'm the MemTest86 developer / maintainer. So obviously bias.
But for what it is worth here's some comments on the differences.
When run on old 32bit systems (pre UEFI BIOS being available) there isn't much difference in them. The code base is similar, the tests are similar and they are both open source.
On newer systems (UEFI + 64bit) there are significant differences. MemTest86 was rewritten for UEFI and 64bit. The main new features in MemTest86 that aren't in MemTest86+ are,
- The code and memory addressing are now 64bit
- DDR4 is supported, including DDR4 XMP
- ECC is supported
- Secure boot is supported, MemTest86 is now code signed by Microsoft
- Multiple languages are supported. Chinese, Spanish, German, etc..
- No longer uses the Linux boot loader (but runs direct in UEFI)
- Row hammer & SIMD instruction testing was added
- Significant speed improvements
- Mouse support and GUI support
- Writing of test reports to disk (there is still no way to save results in MemTest86+)
- RAM SPD support for new CPUs & Chipsets
But as pointed out MemTest86 V6 is no longer GPL. But there is still a comprehensive free (no strings attached) version available. FOSS is nice, but having something that is current and works is important as well.
For now, stay with memtest86.
Ran memtest86 and found no errors.
Ran prime95 and found no errors.
Should I assume that the motherboard, or some other hardware is faulty?
Edit:Can a faulty motherboard corrupt the instruction pointer?The last appcrash I received did not have a corrupted instruction pointer, but was still memory corruption. Can a faulty motherboard corrupt memory then, or is there something else I should look closely at?
Problem seems to be getting worse now, though I'm surprised no BSOD's have shown up.
Last edited by silentreaper; 31 Jul 2016 at 06:09.
Yes, the motherboard can cause almost any kind of error, including memory corruption. Defective memory ports on the board have been encountered before.
Try removing 1/2 of your RAM, leaving 8GB. Run that way for awhile and see what happens. If okay it's possible the removed RAM is defective. Sw2ap removed and installed RAM and test some more. See if you can narrow things down this way.
Following shows where the single module should be installed.
Attachment 93242
I tried two sticks of RAM, in A1/B1, and still got error c0000005. Took them out, placed the other two in, same error. However, upon accidental restart, I received Q-Code error 55: Memory not installed. I haven't touched them, and they are correctly in the motherboard.
If it helps diagnosis somehow, I did have a problem turning on the computer two weeks ago, trying to figure out another problem. I couldn't even boot the computer until I hit the Clear CMOS button.
Two modules should go in A2 and B2 according to the diagram above.
Single stick in A2, two in A2 and B2.
Did the same for A2/B2, still got errors, minus the Q-Code error
Quote