Asus M2N-E vs. the MSI K9N SLI Platinum Motherboard

I purchased an Asus M2N-E back in March of 2007, and it worked like a charm for about 6 weeks. However one evening my nightly memtester cron job reported an error and after further investigation with some disk-based memory testers (scripts I wrote myself) it appeared that there was indeed a memory fault on my system. Interestingly though, after rebooting and running memtest86+ it took hours to see an error, whereas running the system 'full-tilt' would result in an error almost immediately*.

I booted from alternate media with different kernels, then I stripped the system clean, removing add-on cards and disabling on-board peripherals trying to determine if any of the components were at fault, yet nothing helped.

I tried running my RAM (high quality OCZ DDR-2) in another system and it worked flawlessly. This made me believe that either the CPU or motherboard was at fault.

After some further investigation via google I discovered that there is indeed a memory problem with the M2N-E.

I have been buying Asus products for many, many years. Of the dozens of motherboards, video, SCSI, etc. cards I have purchased from them, only two video cards had to be RMA'd. So overall I was quite happy with Asus. I think it is wrong for Asus to allow a product with a known issue continue to be sold on the market.

So instead of buying another Asus motherboard to replace my M2N-E, I purchased an MSI K9N SLI Platinum, and it has worked like a charm. I read some good reviews and liked the feature set. If this MSI motherboard keeps on working well then the next system I setup will contain MSI hardware and not Asus products.

UPDATE: It was just over a month and the MSI K9N SLI Platinum system had worked flawlessly, however the OCZ RAM started to display memory errors (found via my nightly memtester runs, and yes the memory was tested in another DDR-2 system for comparison). As a result I have RMA'd the OCZ RAM and purchased an identical set as a temporary replacement, which is still working without error. Got back the OCZ RAM from RMA (took only a week - I have to say that OCZ has a great trouble ticket and RMA system, with very friendly staff!) however the replacement RAM had an even higher number of errors! I am in the process of getting this second set RMA'd, luckily they offered to cover the cost of shipping.

UPDATE: The Asus M2N-E returned from RMA and I tested it with memtest86+ and memtester for a full 24 hours with no errors. It will be interesting to see how it fairs after a month or two.

UPDATE: The Asus M2N-E that was RMA'd has worked flawlessly for 6 months now!

*In my expert opinion (as an Electrical Engineer who designs embedded systems for a living) I believe that memtest86+ found fewer failures because the peripherals are inactive during memtest86+ (only the CPU and memory system are really being used). When the entire system is running, especially the disk, there are two major side-effects. First, background disk/memory DMAs are occurring that can drive much more data than the CPU can generate on its own (not as applicable with a 2 GHz CPU clock, yet it does contribute). Second, there is a much larger drain on the power generation hardware due to the extra work being done by the CPU chipset and memory. I am guessing that my disk-based memory test produced the largest number of memory failures because of a failure in the components that perform the DC-DC conversion that power the chipset and RAM/CPU. In fact I wouldn't be surprised to discover that this is the result of another bad batch of capacitors! Just a side note: both motherboards contain the NVIDIA nForce 570 chipset (the MSI uses the SLI version).