OCEAN v3.2.0

v. 3.2.0

Minor features

1. Updates to the screening defaults. Should still be fast, but closer
   to convergence
2. Broadening approximation for XAS is no longer the default (column 2) in
   outputs. The unbroadened spectrum (column 3) was already recommmended.
3. Updates to the Pseudo-Dojo potentials to add in f channels to the
   following elements: Ba, Bi, I, Pb, Po, Rb, Rn, S, Te, Tl, and Xe. This is
   in line with the v0.5 updates. They've been applied to LDA and PBEsol as
   well as PBE.

v. 3.1.3

Bug fixes

1. Crashing bug in RIXS of 2p when the underlying DFT calc had spin=1
2. Fixed for ABINIT compatibility
3. Small fixes for GW corrections

v. 3.1.2

Minor features

1. Reduction in the memory required for screening calculations. Could be a
   substantial improvement for large systems.

Bug fixes

1. Small, odd number k-point grids might have had slightly incorrect x-ray
   spectra due to an error in how the k-point sampling was unfolded into a
   supercell.
2. Gamma-point selection for screening calculations was not being triggered.

OCEAN v3.1.1

v. 3.1.1

OCEAN v3.1.0

v. 3.1.0

Bug fixes

1. The core-level shift was including the effect of V_{xc}. For self-energy
   corrections, V_{xc} should be removed. Initial testing has shown that the
   relative shifts between sites w/ or w/o V_{xc} are largely the same, but
   in absolute terms there will be a large change (10s of eVs).

OCEAN v3.0.5

Merge pull request #320 from times-software/develop

Update VersionNumber

OCEAN 3.0.4

v. 3.0.4

Minor features

1. The valence BSE can now self-consistently determine the static dielectric
   constant by setting bse.val.epsilon_threshold to be > 0.
2. When using QuantumESPRESSO, the static dielctric constant of insulating
   systems can be determined using density-functional perturbation theory.
   This is attempted automatically if there is no dielectric constant set in
   the input file. Systems run as metals will use a default value.

Bug fixes

1. Improved compatibility with older versions of QE.
2. Problems in the valence BSE with spin systems for both valence and RIXS.

OCEAN 3.0.3

v. 3.0.3

Bug fixes

1. A small, odd number for the first k-point dimension combined with a
   different number of x-points in the first two dimensions could lead to
   problems.
2. Some x-mesh grids could lead to failures in processing the density for
   valence calculations

OCEAN 3.0.2

v. 3.0.2

Minor features

1. MPSE support added back. The many-pole self-energy model from AI2NBSE has
   been re-enabled. The scripts and source are in POST/MPSE. This is a post-
   processing step for after valence UV/optical calculations.

Bug fixes

1. Screening had small bug. Mostly didn't seem to change spectra, but
   induced potential could change slightly from run to run.

OCEAN 3.0.0

Merge pull request #289 from times-software/develop

Develop

OCEAN 2.9.7

v. 2.9.7

This is the pre-release/test run for 3.0

Major Features

1. Support for D. R. Hamann's Optimized Norm-conserving Vanderbilt pseudopotentials
   1. Create your own
   2. Use built in database from PseudoDoJo collection
2. Improvements to parallelism
   1. Screening calculation is faster
   2. Prep stage is faster too
3. The DFT calculations for the BSE are split in two whenever there is finite-q

Minor Features

1. OPF calculation should be slightly more robust, but noticeable differences are unlikely
2. Large finite-q valence calculations should be working again

OCEAN 2.9.6.3

v. 2.9.6

Major Features

1. Support for D. R. Hamann's Optimized Norm-conserving Vanderbilt pseudopotentials
2. Improvements to parallelism
   1. Screening calculation is faster
   2. Prep stage is faster too
3. The DFT calculations for the BSE are split in two whenever there is finite-q

Minor Features

1. OPF calculation should be slightly more robust, but noticeable differences are unlikely
2. Large finite-q valence calculations should be working again