David Bowler

This happens when the number of atoms per process is small (1-2 atoms/process). Typically the partitioner has to use an excessively large number of partitions to distribute the atoms as...

Add tutorials

2

We will develop and add tutorials for Conquest.

Problem with MSSF optimisation and full cell optimisation (option 3)

1

When performing full cell/ion optimisation (using `AtomMove.OptCellMethod 3`) with MSSF and optimising the basis set, after the initial ground state search there is some error in how the density is...

Implement band density/STM post-processing with MSSF

1

The post-processing utility does not have the capability to deal with MSSF (and this will require extra output from Conquest as well) for band density and STM simulation.

We need to check the details of the local potential calculations when not using the neutral atom for ionic (i.e. non-neutral) pseudopotentials. In particular, whether `zval` is the correct value...

Analysis of electronic structure for large systems

2

We will enable the analysis of electronic structure from large-scale linear scaling calculations via the Sakurai-Sugiura method. This is likely to have two stages: 1. Output of matrices for reading...

The Hamann ONCVPSP code has a number of issues that we would like to work around: 1. It seems to require partial core corrections (PCC) in many cases (it produces...

We need to improve the MSSF optimisation: 1. Change the SCF cycle so that the LFD is done at each iteration 2. Add option for efficient (non-SCF) PAO optimisation after...

We envisage OSSF being used mainly with O(N) (and MSSF with diagonalisation) but it is still not clear how to optimise the OSSF most efficiently. Some notes: 1. We know...

Revisit blip function optimisation to improve efficiency and speed.