~nickpapior/siesta/4.0-trs : changes

~nickpapior/siesta/4.0-trs » Changes from revision 561

From Revision 561 to 542

Rev	Summary	Authors	Date
561	Patched diag2[gk] to conform to sign-conventions in 4.1 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.	Nick Papior	6 years ago
560	Fixed documentation for TRS Fixed documentation for TRS	Nick Papior	6 years ago
559	Merged 4.0 into trs Merged 4.0 into trs	Nick Papior	6 years ago
558	Stream-lined NC mulliken output with 4.1 Stream-lined NC mulliken output with 4.1	Nick Papior	6 years ago
557	Merged 4.0 Merged 4.0	Nick Papior	6 years ago
556	Further cleaned interface to clear out an inconsistency in t... Further cleaned interface to clear out an inconsistency in the code	Nick Papior	6 years ago
555	Removed forced TRS for non-colinear spin 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.	Nick Papior	6 years ago
554	Removed all references to __FILE__ to enable conda-builds, f... Removed all references to __FILE__ to enable conda-builds, fixes lp:1746096	Nick Papior	6 years ago
553	Possible bugfix for paralleloverk with k-grid in SCF, and Ga... 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.	Nick Papior	6 years ago
552	Added a test for PDOS using the Gamma-only for kpoint calcul... Added a test for PDOS using the Gamma-only for kpoint calculations	Nick Papior	6 years ago
551	Fixed spin-polarized pdos.xml writing Fixed spin-polarized pdos.xml writing Wrong code in r550.	Nick Papior	6 years ago
550	Extended PDOS files with more metadata Extended PDOS files with more metadata The PDOS xml output now also adds: - Z: atomic number - P: true/false for polarization orbital Also reduced temporary arrays and merged to loops. Renamed pdos.xml file to siesta.PDOS.xml to make it system-label dependent. I have checked the utilities and there is no problems with those.	Nick Papior	6 years ago
549	Increased length of int->str in wxml Increased length of int->str in wxml This fixes a potential problem when a very large integer tries to get represented by the str_integer routine in wxml. This simply increases the format to allow a wider field.	Nick Papior	6 years ago
548	Final edits for read/write of floating orbitals Final edits for read/write of floating orbitals Also fixed an inconsistency in the interpolation routine name.	Nick Papior	6 years ago
547	Fixed writing/reading ion.nc files for ghost-atoms Fixed writing/reading ion.nc files for ghost-atoms Now siesta writes out ion.nc files for ghost-atoms, in these cases will the number of projections be 0. Secondly, a bug has been fixed when trying to read an ion.nc file that was not present. This fixes lp:1738425	Nick Papior	6 years ago
546	Bug in writing a Bessel function ion/ion.xml file Bug in writing a Bessel function ion/ion.xml file Since the Bessel basis does not contain any pseudo-potential siesta would write binary stuff in the ion/ion.xml files which was preventing it from being read elsewhere. I have added a pseudo-potential initializer which defaults lots of values to the pseudo-potential (could be necessary for Bessel and other exotic pseudos. Also ensured lots of type arrays to be nullified upon creation, which should help in the aid when debugging.	Nick Papior	6 years ago
545	Bug re-read basis functions with no potentials to interpolat... Bug re-read basis functions with no potentials to interpolate (fix lp:1736455) When re-reading a basis from a ghost or Bessel orbital then many of the quantities are 0 (Vna, VCore, etc.). Since 4.0 the interpolation routine prevented these from the spline interpolation because there is a check for monotonic arrays (all 0). This commit fixes this by simply checking for erroneous input, and if so it will set dy2/dx2 to 0. This is rather arbitrary but at least initializes the dy2/dx2 to some values. Secondly, a possible memory leak in interpolation is corrected by issuing a fixed clean_spline call.	Nick Papior	6 years ago
544	Merged fix lp:1728281 Merged fix lp:1728281	Nick Papior	6 years ago
543	Added compiler version information Added compiler version information Now the Makefile will automatically add the relevant compiler information such that one can more easily determine the compiler version used.	Nick Papior	6 years ago
542	Fixed monkhorst-pack displacement, lp:1721479 Fixed monkhorst-pack displacement, lp:1721479 This fixes an uncommon bug where the displacement was more than 1. Now all displacements are translated into the principal BZ. I have tested this by running h_chain with 0.5/7 and 1.5/7 with exactly the same results.	Nick Papior	6 years ago

Older »