I understand that the MIG is provided by Xilinx and I wouldn't (and didn't) ask you to resolve their problems. I'm talking about problems out of the box with the Opal Kelly RAMTester example on the XEM6310 board.
The first error is the fifo was built for a different spartan part. The .xco file is there so that is easy to recreate. Sort of makes you think that this has never been tested with the XEM6310.
The MIG is much harder. There are many more parameters and the ports have different names with the current version of MIG. I'm dissappointed that an up to date example isn't provided by OK to get started with. Like I said, this release of ISE and MIG are 3 years old.