T. Daniel Crawford

Results 18 comments of T. Daniel Crawford

I agree that it should warn the user more clearly, but the lack-of-exit behavior is intentional. It's an attempt to salvage useful data rather than simply quitting and giving nothing...

Please note that `libdiis` did not exist when the DIIS code was written in the `cc` modules. Convergence difficulties in `cclambda` are not due to DIIS difficulties but due to...

I agree that `cclambda` should honor the user's `r_convergence` choice, but the default settings we chose are not to avoid breaking tests but because tighter settings are wholly unnecessary (and...

FYI, we did do a brief convergence study, which is how we came up with the 1e-4 criterion.

In a spin-adapted RHF formalism, these two terms are distinct. Antisymmetry of the integrals and HBAR matrix elements only applies in a spin-orbital formulation.

Can you elaborate on what you mean by "the main scope of the function is quite polluted"?

> Certainly. Polluted might have been too harsh, cluttered may be a more appropriate word for it. > > All variables are currently declared at the beginning of the function....

Just to confirm: these are almost all just removal of the C-based wrapper functions, correct?

Yes, I believe the global PSIO object was to created to minimize the changes needed to older modules that worked with the pre-OOP version of the library.