Sounds like most everyone has got this figured out, but I just thought I'd add my two cents to the mix in an effort to help document this issue:
Enlight 7250 w/350 w psu
MSI K9N Platinum (not sli)
AMD Athlon64 x2 4200
2 GB Kingston PC5300 667 RAM (1GB x 2)
ASUS 6600 Video
Western Digital 320GB 7200 rpm
Lite-On 16x DVD+-RW x2
NEC floppy
Cyrius, first of all, I had actually done exactly what you did, and bought Kingston memory with this computer I am building for a customer. Kingston is very compatible with MSI. One MSI tech told me once that Kingston is right across the street from them in fact.
When I put the system together 2 weeks ago, it booted up but would not load Windows. I thought it was the Windows disk at first, but I then thought it might be the memory so, I took out 1 stick and got it to load Windows with some trouble. I then tested both sticks of RAM with memtest86+ v1.65 and both were "bad". I RMA'd the RAM back to Newegg, and after a week, got the replacements. I installed these and lo and behold, memtest86+ tested these out as being "bad" also. Lots of errors in a short time. Newegg was very nice and actually next day shipped out another set of RAM as I RMA'd back the 2nd "bad" set to them for refund. When I got the third set of RAM memtest86+ tested these out as "bad" also. That's when I knew something had to be wrong with the motherboard/chipset. I did a little bit of searching and found out that this seems to be a pretty common issue as you all have noted, with the new AM2 socket and the new nVidia chipset, along with ddr2 RAM. The Kingston memory which I had installed is rated to run at 1.8v and that is what the MSI board defaults to. I had just figured naturally that the motherboard/chipset wasn't at fault because I was running the RAM at its rated speed and voltage. After reading a few posts I tried underclocking the RAM to 400 to be extra safe and it worked perfectly. No memtest errors. Then I tried putting it back to 667 and the errors were back, so I raised the voltage incrementally from 1.8 to 1.85 to 1.9 to 1.95 to 2.0 until it hit 2.05v at 667 and it works perfectly now, with no memory errors from memtest86+. Each time I raised the voltage I got less and less memtest86+ errors and Windows became more and more stable. 2.05v happened to be the "sweet spot" for the voltage on my particular machine.
As a side note, Cyrius, whenever I had those memtest+ errors before I raised the voltage to the RAM I could usually boot to Windows, but the OS was very unstable and frequently rebooted randomly. This may be what you are experiencing if you haven't tried raising the voltage to your Kingston RAM yet.
As others have mentioned, anytime new technology like AM2 comes out, all of us should expect to run into problems like this until BIOS revisions, drivers, and the technology in general has had the time to mature in the market. I'd give it 2 months and we'll probably see all board manufacturers and memory manufacturers to have fixed this with revisions to their products and BIOS updates etc. This isn't JUST an MSI + OCZ problem.
Also, in defense of MSI, I've actually had pretty good luck with MSI tech support, and I pretty much exclusively use their boards. Have built or supervised the building of 300+ MSI based systems from 2002 to present.
Please reference the following link where the third reviewer down had the same problem with Corsair Memory and fixed it with Voltages/Timings:
http://www.newegg.com/Product/CustratingReview.asp?Item=N82E16820145098
I respect this reviewers remarks about the high RMA rate and how it is not bad parts, but parts not playing well together . . . yet.
I'm hoping for BIOS revisions soon. After all, we shouldn't HAVE to raise the voltage on our memory or mess with timings to get our systems to work at their rated speeds. This is very frustrating of course. Raising voltage/tweaking timings should be for overclocking only, not for JUST getting your system to work.
Hope all this helps everyone concerned.