The tutorial on the map class (http://eigen.tuxfamily.org/api/TutorialMapClass.html) lists two ways of specifying that the data array is laid out in row-major format:

1) map the data to a row-major matrix
2) map the data to a col-major matrix, but specify that the outer stride should be one and the inner stride equal to the number of columns.

Are these methods equivalent, or are there scenarios when one would be more efficient than the other?

The first method is preferred. InnerStride != 1 is not recommended performance-wise.