RM mode and its effects on RRM determinant of AM1 matrix

[Gene Wright]

RM Mode       Determinant

-------   ---------------------

  0       0.99999 711935 38228

  1       0.99999 720675 30552

  2       0.99999 709507 09280

  3       1.00000 009178 43880

  4       0.99999 711925 74617

  5       1.00000 895891 46320

  6       0.99999 718613 18536

Looks like RM 3 is closest to 1. Fairly significantly closer, it appears.

[Paul Dale]

RM 3 is round half up. It being closer is most likely a fluke.

This does show that the algorithm isn't overly sensitive to rounding mode which is goodness.

- Pauli