memtest+ problem

[mushroom3]

Hi, when i first got my ram, i popped both sticks in and ram memtest for 24 hours. After about 22 hours, it had 3 errors, but the list thats highlighted in red only showed 1 thing

So I take both stick out, ran memtest for 12 hours with each, with no error showing

Afterwards, i test both stick together again for 30 hours this time, yet no error is showing

My question is, was the first incident a fluke? What could have caused it?

 

[montag451]

With your first test/errors - were all the errors located in the same part of storage?
Are you sure you didn't fiddle around with voltages/timings before doing the second test?

If not, just remember that you have an amazing piece of electronics that you are testing - its got tiny pathways that are measured in nanometres, and does things measured in nanoseconds.
Things that might affect a result:
A tiny piece of dust with a charge on it, a tiny little itty bitty power surge, or a cosmic ray, or a stray gamma ray etc etc -

 

[RebateMonger]

Product failure is a statistical process. Even "perfect" memory CAN have errors, due to cosmic ray hits. An older Intel paper estimated the error rate (for 1 GB of memory as I recall), at around one error per month. Again, I'm taking this from memory...but you get the general idea.

I had something like your experience happen to me with some Rosewill DDR memory. A single error that I could never duplicate, even with substantial testing. I ended up keeping that memory. But I don't think I ever used the PC that I put it into. Your choice.