On Wed, May 12, 2004 at 02:57:59PM -0700, Aaron Trumm wrote: I used memtest86 recently and definitely saw errors. There was an error count in the error count column. In the main lower portion of the screen a scrolling buffer showed the addresses of the bad RAM in hex and MB along with the pattern expected and the pattern read and which bits of the address had the errors. When I removed modules and re-ran the test I was able to figure out which stick was bad. When it wasn't in the box there were no errors, when it was in a different slot the addresses of the errors changed in accordance with the size of the sticks (3x512MB). i.e. at first the errors were at around 109MB while the faulty module was in the first slot and later when it was in the third slot the errors were around 1133MB (109+1024). You would definitely know if it had found something. If you leave it running it loops through the tests continuously, but as far as I could tell it kept a running total of errors for all passes. Did you run the extended tests? These are tests 8-12 (i think ? ...) and take many hours. The docs say there is a very low probability of these finding problems the standard set doesn't find. For me they didn't tell me anything I didn't already know, they just found more errors on the same module. So, maybe there is a small chance those would find something more for you. <shrug> Regarding differnt memory speeds, that was one test I wanted to run. But, I haven't had time, yet. I'm wondering if Asus lied when they said this board could handle 1.5GB of PC2700 or if crucial might have mislabled a PC2100 module as PC2700 or something. My thought was to put the bad one back in and set the RAM speed lower in the BIOS and run the tests again. -Eric Rz.