Knytestorme
Member
Before I start I should say I think I have solved the issue but wanted to get some confirmation that my thinking is correct.
I recently built a new 2011-3 PC using an MSI Gaming 9 motherboard with 64Gb of RAM. One of the sets of RAM were faulty and had to be warrantied and while I had 32Gb in everything was fine, after I got the other RAM back I put the first set in DIMM1-4 and the repaired set in DIMM 5-8 so I could remember which was which with both sets being RipJaw 4 with same speed and timings.
This weekend I sat down and went through crash dumps that kept saying it was a driver issue so I reformatted with Win10, installed just the chipset and nVidia drivers, removed all the USB devices except KB/M and unplugged all drives except the boot SSD and let it run but still got crashes.
As a last resort I ran Memtest86 for about 15 hours and saw it got errors on test 4, 5 and 8 so I pulled out the RAM that had been warrantied and placed the other set back in DIMM 1-3-5-7 and ran Memtest through the 10 tests and got no errors. I then powered down, put the second set in DIMM 2-4-6-8 and ran through the 10 tests for about 6 hours and got no errors and machine has still not crashed after 2 days in windows with drives reconnected.
I ~think~ the issue was that I didn't have the matched sets in the alternating slots like I do now and that was the issue due to being quad channel even though both sets are the same manufacturer, speed, timings etc and that instead having them in 1-4 and 5-8 was causing some sort of conflict between the modules (eg 1-3 being from one set and 5-7 being the other set) so I'd thought I'd ask here to see if that is a reasonable explanation for the problem or not.
I recently built a new 2011-3 PC using an MSI Gaming 9 motherboard with 64Gb of RAM. One of the sets of RAM were faulty and had to be warrantied and while I had 32Gb in everything was fine, after I got the other RAM back I put the first set in DIMM1-4 and the repaired set in DIMM 5-8 so I could remember which was which with both sets being RipJaw 4 with same speed and timings.
This weekend I sat down and went through crash dumps that kept saying it was a driver issue so I reformatted with Win10, installed just the chipset and nVidia drivers, removed all the USB devices except KB/M and unplugged all drives except the boot SSD and let it run but still got crashes.
As a last resort I ran Memtest86 for about 15 hours and saw it got errors on test 4, 5 and 8 so I pulled out the RAM that had been warrantied and placed the other set back in DIMM 1-3-5-7 and ran Memtest through the 10 tests and got no errors. I then powered down, put the second set in DIMM 2-4-6-8 and ran through the 10 tests for about 6 hours and got no errors and machine has still not crashed after 2 days in windows with drives reconnected.
I ~think~ the issue was that I didn't have the matched sets in the alternating slots like I do now and that was the issue due to being quad channel even though both sets are the same manufacturer, speed, timings etc and that instead having them in 1-4 and 5-8 was causing some sort of conflict between the modules (eg 1-3 being from one set and 5-7 being the other set) so I'd thought I'd ask here to see if that is a reasonable explanation for the problem or not.
