Are you also zeroing out the diagonal for FMI in the UHF case? I looks like only the Fmi matrix is cleared there.

Yes, on lines 75-77. Look for the little patch of green between the red.

Are you also zeroing out the diagonal for FMI in the UHF case? I looks like only the Fmi matrix is cleared there.

Yes, on lines 75-77. Look for the little patch of green between the red.

OK, now I see what you're doing. My only suggestion is that you modify the comment under RHF to read:

// This is because the amplitude update in RHF uses the full residual while the ROHF and UHF updates separate out the diagonal contributions from the Fock matrix [cf. Eqs. (1) and (2) of Stanton et al., J. Chem. Phys. 94, 4334-4345 (1991)].

Mostly adopted : I changed a few words to specify that these are true of Psi's implementations, not necessarily all implementations.

Thanks for the review! I will always accept specific journal references.