Skip to content

Latest commit

 

History

History
56 lines (39 loc) · 1.86 KB

RELEASE.md

File metadata and controls

56 lines (39 loc) · 1.86 KB

MOLGW: Release Notes

What's new in version 1.H

Overview

Bug fixes

Changes affecting the compilation

  • MOLGW is now compatible with LIBINT versions 2.4.x
  • Possibility to compile MOLGW with LIBINT having the one-body integrals, but not the gradients. Use the preprocessor flags -DHAVE_LIBINT_ONEBODY and/or -DHAVE_LIBINT_GRADIENTS

What's new in version 1.G

Overview

Bug fixes, cleaning, and speed-up.

Changes affecting the compilation

  • Still not possible to link with the versions 4 of LIBXC, due to missing functions on their side

Changes affecting the usage

  • The default value for read_restart is now set to 'no'
  • Speed-up in the semi-local DFT Vxc calculations thanks to the use of batches

What's new in version 1.F

Overview

A few bugs specific to the recent versions of the Intel compilers have been fixed.

What's new in version 1.E

Overview

Bug fix with respect to previous version for high angular momenta (L>5) Considerable speed-up in the diagonalization of the RPA matrix thanks to the use of PDSYEVR instead of PDSYEV

What's new in version 1.D

Overview

Simple bug fix with respect to previous version

What's new in version 1.C

Overview

This release makes better use of the latest version of LIBINT ( >= v2.2.0). Together with some internal refactoring of the code and small bug fixes.

Changes affecting the compilation

  • Linking with a recent version of LIBINT (>=v2.2.0) is compulsory
  • Compilation flag -DHAVE_LIBINT_ONEBODY activates the use of LIBINT one-body integrals and integral gradients
  • Use of libtool is no longer necessary

Changes affecting the usage

  • Input variable no_4center has been removed. From now on, the calculations use or do not use resolution-of-identity from the beginning to end.