It because the memory you added is mismatched to whats already in the server.
For the R610 to be in "optimal" configuration, not show that error message, and get through POST without prompting F1 to continue, all of the ram sticks need to be the EXACTLY the same.
Same size, same speed, same latency, same number of chips per side, matching rows and columns etc.
What is really going on, once you read through the fine print, is that the "optimal" memory configuration for the R610 is the same as a home PC's "dual channel" memory mode. I.e. the two memory banks both run in parallel to effectively double the ram bandwidth. However dual channel only works if all your memory is the same.
So you have three choices:
1. Take back your RAM and get stuff that is the same as whats currently in your server.
2. Ditch the servers current RAM, and buy 12 new sticks that are all exactly the same, from the same manufacturer, and put those in.
3. Tell the BIOS to ignore F1 to continue POST error messages, and just accept that your memory bandwidth will be a little lower. <--- This is what I did. (Haven't noticed a difference in performance except in memory bandwidth tests ... meh).
As for why Dell decided that non optimal memory configurations should be considered critical enough to stop the server from getting through POST?
... Well I can only assume they are massive idiots. Image may be NSFW.
Clik here to view.
