lp:~nickpapior/siesta/4.0-trs

Branch merges

Propose for merging

No branches dependent on this one.

Merged into lp:siesta/4.0 at revision 564

Nick Papior: Needs Resubmitting on 2018-04-12

Ramon Cuadrado (community): Needs Information on 2018-02-01

Alberto Garcia: Pending requested 2018-01-31

Diff: 466 lines (+97/-141)

8 files modified

Docs/siesta.tex (+5/-4)
Src/Makefile (+4/-4)
Src/diag2g.F (+23/-26)
Src/diag2k.F (+50/-95)
Src/kpoint_grid.F90 (+7/-5)
Src/kpoint_pdos.F90 (+7/-5)
Src/mulliken.F (+0/-1)
version.info (+1/-1)

api

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

561. By Nick Papior on 2018-04-12

Patched diag2[gk] to conform to sign-conventions in 4.1

This also lead to the discovery of a sign error in the phases.
So now everything in 4.0 is the same as 4.1 (for non-collinear).

I have checked fe_clust_noncollinear and FeO.
Both yield, exactly the same output before and after this commit.

560. By Nick Papior on 2018-04-12

Fixed documentation for TRS

559. By Nick Papior on 2018-04-12

Merged 4.0 into trs

558. By Nick Papior on 2018-04-11

Stream-lined NC mulliken output with 4.1

557. By Nick Papior on 2018-04-11

Merged 4.0

556. By Nick Papior on 2018-01-31

Further cleaned interface to clear out an inconsistency in the code

555. By Nick Papior on 2018-01-31

Removed forced TRS for non-colinear spin

Since 4.1 implements no time-reversal symmetry we deemed it
necessary to remove TRS. This is also based on comments on the AbINIT code.

The generation of the DM does not use the TRS requirement of the wavefunctions.
I.e. there are no complex-conjugation averages (diag2k).

Also the k-point generation now defaults to *not* use TRS. for non-colinear spin.

Also there were some low-hanging fruits for memory optimizations regarding the
diag2k routine. I have cut the memory requirement almost in halve.

554. By Nick Papior on 2018-01-30

Removed all references to __FILE__ to enable conda-builds, fixes lp:1746096

553. By Nick Papior on 2018-01-04

Possible bugfix for paralleloverk with k-grid in SCF, and Gamma in PDOS

This is strictly not required in 4.0, but in the later versions this
is necessary.
This commit fixes the inconsistency of updating the "global" parallel-overk.

552. By Nick Papior on 2018-01-04

Added a test for PDOS using the Gamma-only for kpoint calculations