GitLab

Note to myself and future generations: 2**p is an divisor of 2**m even if m % p! = 0. situation: managed to fix the calculation of the eigenvalues (checked against the values produced by the ref versions that build the complete H). overhault pt4/?