I've been tackling this problem for quite a while now, and I tried several things, but I'll try to explain the best I can. Bear with me.
Just under a year ago I built a new computer using, among other components, an Asus P5N32-E SLI motherboard and four 1GB Corsair Dominator 800MHz DDR2 DIMMs (CM2X1024-6400C4D), for a total of 4 GB of RAM. I installed Vista Ultimate 32-bit.
It seemed to work fine except for the occasional crash, which I attributed to Vista, the BIOS and the device drivers being new. However, when I tried to install SP1 and failed, a KB article recommended that I run Vista's memory diagnostic tool. I did, and it found memory errors. Crashes were becoming more often, so I downloaded memtest86+, booted to it, and after a couple of tests, I found the faulty module. I removed it and the module it was "paired" with, leaving 2 GB in dual-channel. I took the opportunity to format and upgrade to Vista Ultimate 64-bit.
Everything ran fine for a few days, and then again crash. Now the situation is as follows:
- With 2 GB of RAM the system is fairly stable, but it will get a BSOD about once a day. As far as I can tell, the error referenced in the BSOD is different every time. BSODs do not appear when doing anything in particular, they may happen when the computer is just idling or whatever.
- With 3 GB using the 3 "good" modules which pass memtest, the system is unusable. Windows will BSOD right after booting (usually PAGE_FAULT_IN_NONPAGED_AREA stop, but not always), and an Ubuntu live CD won't boot (different error each time).
- With 4 GB using the four modules, including the faulty one, the system is much more stable than with 3 GB (that is, adding a faulty module increases stability!), but it will still crash about as often as with 2 GB.
I tried memtesting all the modules, in different combinations and different slots, and all pass except for the known faulty one. HDD is discarded as a cause because Ubuntu live CD doesn't use it. The power supply was changed and problems remain the same. All components are running at stock speeds and were never overclocked, and the system is nicely cooled. Since both Ubuntu CD and Windows have the same problems, it is definitely hardware.
I'm running the latest BIOS. I upgraded it mostly because with the factory version, the computer would often not awaken from sleeping mode correctly, and I had to reboot. I did try different BIOS settings. Until I started dealing with this issue, most settings were set to "auto". When I started diagnosing this, I noticed that the memory timigs were incorrectly detected (5-5-5-18 instead of the modules' rated 4-4-4-12), so I set it manually, and I set the memory voltage to the recommended 2.1v (instead of auto) while I was at it. The problem seems to be the same with both settings, except that with 5-5-5-18 the system was stable for a little longer when using 3 GB, although it still crashed after a short while.
I'm going nuts with this. I know I will have to replace some hardware, but is it the memory? Is it the motherboard? How can I find out?
You might try removing the bad DIMM and it's matched DIMM and run the two good DIMMS. Set the RAM voltage to 2.2v. and be sure to bump the Northbridge voltage to 1.5v. Try 5-5-5-15 and run memtest. If you are error free you may want to adjust the CAS latency's lower. If the problem still exists, try some different RAM.
If your system's front panel speaker header is connected to a working speaker and is capable of giving BIOS beep codes, you can remove ALL THE RAM and boot the system. If the MB fails to beep indicating NO RAM in the system you have a bad MB.
You might try a RAM configurator (memory advisor tool) like this and buy RAM that is guaranteed to work in your MB:
You are about to answer a thread that has been inactive for more than 6 months. If you still wish to proceed, please ensure that your posting is original and does not duplicate or overlap any prior responses to this thread.