~dikan.va/siesta/thtr-dm-deriv-output : changes

~dikan.va/siesta/thtr-dm-deriv-output » Changes from revision 792

From Revision 783 to 764

Rev	Summary	Authors	Date
783	KS_flucs_procs refactoring KS_flucs_procs refactoring Explicit allocations of psi-arrays in `compute_Jks`. H -> H + eo in `compute_Jks`	Vladimir Dikan	4 years ago
782	Jks component of heat flux Jks component of heat flux Had to add storage for R-matrix to `m_ks_flux_data` Need to organize `m_ks_flux_procs`	Vladimir Dikan	4 years ago
781	Data saved from base step is used to compute psi-s in virtua... Data saved from base step is used to compute psi-s in virtual step	Vladimir Dikan	4 years ago
780	rearrange virtual step data & subroutines rearrange virtual step data & subroutines In order to resolve circular imports, virtual_step_md is split into modules for interstitial system data storage and procedures. Trivial subriutins for saving/restoring parameters of system in question between Base and Virtual steps (forces, positions, velocities) are moved out of m_virtual_step_procs and put under siesta program contains-statement for clarity.	Vladimir Dikan	4 years ago
779	psi_dot_c: all three multiplication rows in place psi_dot_c: all three multiplication rows in place	Vladimir Dikan	5 years ago
778	optimized loop for matmul scheme in psi_dot_c optimized loop for matmul scheme in psi_dot_c	Vladimir Dikan	5 years ago
777	2nd row of psi_dot_c comp. scheme is computed 2nd row of psi_dot_c comp. scheme is computed And is compared to the "functional" approach made for self-test. The grouping of matrix multiplications differs, yet the results are the same (associativity). Further TODOs: - the rest of the contractions - computation of Jks - moving of the "test" module to a separate branch, use it for references	Vladimir Dikan	5 years ago
776	Fixed bug with deletion of gradS between MD steps Fixed bug with deletion of gradS between MD steps	Vladimir Dikan	5 years ago
775	cleanup: removed standalone compute_psi_hat_c cleanup: removed standalone compute_psi_hat_c	Vladimir Dikan	5 years ago
774	Init test for psi_dot_c computation in ks_flux_procs Init test for psi_dot_c computation in ks_flux_procs	Vladimir Dikan	5 years ago
773	Refactoring: KS flux data and procs in separate modules Refactoring: KS flux data and procs in separate modules Code defined in compute_psi_hat_c.f90 was split between m_ks_flux_data and m_ks_flux_procs. Made some order in fdf options. Test flux procedure calls work.	Vladimir Dikan	5 years ago
772	overlap gradient computed on BASE step overlap gradient computed on BASE step	Vladimir Dikan	5 years ago
771	siesta.F: two-step MD loop expansion siesta.F: two-step MD loop expansion	Vladimir Dikan	5 years ago
770	cosmetic changes: calls to size() and re_alloc() cosmetic changes: calls to size() and re_alloc()	Vladimir Dikan	5 years ago
769	renamed Rho_deriv to Dscf_deriv renamed Rho_deriv to Dscf_deriv	Vladimir Dikan	5 years ago
768	Overlap_Gradient stored for Jks flux component Overlap_Gradient stored for Jks flux component	Vladimir Dikan	5 years ago
767	finalized merge of 'virtual_step_md' into 'edens' finalized merge of 'virtual_step_md' into 'edens'	Vladimir Dikan	5 years ago
766	Fix order of slots in Ri in rmatrix Fix order of slots in Ri in rmatrix	Alberto Garcia	5 years ago
765	First installment of computation of R_mu_nu elements for the... First installment of computation of R_mu_nu elements for thermal transport (Work in progress) routine rmatrix in module m_rmatrix in rmatrix.f is modelled after 'overlap'. It computes Rmat, wich is a pointer to the value section of the dSp2D object Rmat_2D. Both live in module sparse_matrices. TODO: -- Use the correct origin for the X,Y,Z matrix elements (matel puts the origin in the second orbital). -- Debug obscure error coming from the spher_harm module. -- Debug parallel operation -- "Deletion" of Rmat_2D in reset_sparse_matrices might need to be wrapped.	Alberto Garcia	5 years ago
764	Merge 4.1:1069-71: lua MPI 0, docs, matel pool lp:1815735 Merge 4.1:1069-71: lua MPI 0, docs, matel pool lp:1815735	Alberto Garcia	5 years ago

Older »