Which pseudopotential formats are most useful and should be standardized?

[prckent]

As pseudopotentials are added and where practical we should standardize available formats. Suggested formats:

• Molpro and GAMESS format Gaussian fits [if practical], and matching basis sets if available.
• CASINO and QMCPACK radial grid based ASCII and XML formats
• UPF format for Quantum Espresso [if practical]

UPF format requires information for the KB projectors which will require additional checking and some agreement with the original author that the conversion is OK. Similarly we might not want to provide Gaussian fits to potentials generated on radial grids unless the fits are highly accurate. Other conversions can be automated on a "per recipe" basis.

What other formats would be useful?

[scemama]

+1 for GAMESS format please, as it is the format we use with Quantum Package.

[dubeckym]

hi, we would appreciate Gaussian sw format.; thanks, m.

[prckent]

The current situation, which is believed to be sufficient. Comment if more are needed:

1. Gaussian formats:
   1. GAMESS
   2. Molpro
   3. NWCHEM
2. Numerical formats:
   1. UPF with KB projectors for plane-wave codes, as projectors are made