- Oct 28, 2000
- 867
- 0
- 0
Last week I upgraded my computer with an i7-3770K, and Intel DZ77GA-70K motherboard, and a Corsair 4x4GB Vengeance Blue LP 1600mhz PC3-12800 memory kit. Everything is at stock. After install everything looked good for a couple days and it passed several runs of IntelBurnTest at the 4GB setting.
Well, since then I've had two BSODs. I ran IntelBurnTest testing 14 GB and it failed at the second run. Then I ran memtest86+ and it failed right off the bat. I pulled sticks one by one until I found the culprit. After keeping that one out memtest86+ was fine.
The computer's been crunching Rosetta@home and looking back at my results, out of 250+ successful work units, 3 had compute errors where they ran fine on another computer. Also, other than those two BSODs, there have been no other crashes or unusual behavior. I've ordered a new memory kit and will return the old one.
Now I'm paranoid about having installed Windows with defective memory. Normally I'd just wipe the drive and start over but I'm extremely busy right now and would rather not have to do this. I'm tempted to take the wait and see approach given the relative infrequency of the obvious glitches the memory caused. What say you, AT? Any experience with this?
Well, since then I've had two BSODs. I ran IntelBurnTest testing 14 GB and it failed at the second run. Then I ran memtest86+ and it failed right off the bat. I pulled sticks one by one until I found the culprit. After keeping that one out memtest86+ was fine.
The computer's been crunching Rosetta@home and looking back at my results, out of 250+ successful work units, 3 had compute errors where they ran fine on another computer. Also, other than those two BSODs, there have been no other crashes or unusual behavior. I've ordered a new memory kit and will return the old one.
Now I'm paranoid about having installed Windows with defective memory. Normally I'd just wipe the drive and start over but I'm extremely busy right now and would rather not have to do this. I'm tempted to take the wait and see approach given the relative infrequency of the obvious glitches the memory caused. What say you, AT? Any experience with this?
