Your Mac Pro with Xeon Processor has substantial additional logic to use Error Correction Code Memory. Eight additional check bits called syndrome bits are stored with each 64-bit word when written to main store, and checked when read back.
The additional syndrome bits provide enough redundant information to perform single error correction (which is performed on-the-fly) and double error detection, which deliberately causes a Kernel panic and halts your Mac Pro before the bad data are propagated into your files.
Each and every memory transaction is watched by this Hardware. Each and every memory error is at least noted, corrected when possible, and the machine is halted when correction is not possible.
A single transient memory error, over a long period of time, is not a cause for concern. It will be corrected and processing will continue at full speed.
When your Mac stops with a Machine Check, uncorrected error, this is a Double-bit error. Your "safety margin" is gone. It will not "get better".
DO NOT IGNORE this valuable information. The DIMM where the error occurred is failing, and will cause you more and more problems.Mac Pro (Early 2009), Mac OS X (10.6.8), & Server, PPC, & AppleTalk Printers