sPuReMD: fix issue with SMALL_BOX_SUPPORT incorrectly computing pairwise atomic distances for periodic self-images. Fix issue with integer vector sum function (used in bond order and torsion interactions).

sPuReMD: remove global variable (array of function pointers for printing bonded interactions during debugging).

sPuReMD: improvements for building in debug mode (add assertion support for run-time checks). Fix issue with Verlet list recomputation logic being incorrect. Fixes to be string comparisons.

sPuReMD: fix issue with certain preconditioners not being initialized for runs using restart files. Fix issue with part of simulation box struct not being initialized for runs using restart files (box->min). Other code clean-up from debugging issues with SMALL_BOX_SUPPORT.

sPuReMD: add safety check for BGF input files (CRYSTX). Remove compiler warnings when enabling DEBUG_FOCUS.

sPuReMD: fix issues with using sizeof with dynamically allocated pointers. Change OpenMP schedule policies for linear algebra code. PuReMD: correct discrepancy with charge matrix construction. Enable dual solver for charges by default.

sPuReMD: refactoring to better align with PuReMD code. PuReMD: backport sPuReMD changes from validation efforts against Fortran ReaxFF code.

sPuReMD: fixes to SAI preconditioner to correctly apply threshold to negative entries and to handle EE/ACKS2 cases when a threshold value of 1.0 is chosen.

sPuReMD: fix bug with diagonal scaling during FG_ICHOL/FG_ILUT producing incorrect factorizations when diagonal entries were negative.

sPuReMD: complete fine-grained Cholesky implementation. TODO: fix numeric breakdowns with water 6540 atom system.

sPuReMD: workarounds for memory issues with ILUT factors. Implement ILU(0) and call when selecting ILUT with a droptol of 0.

sPuReMD: fix issue with hbonds list indices not being reset correctly. Small algebraic rearrangements of valence angle calculations to match Fortran ReaxFF code (no numeric changes). Fix issue with rvec_isZero being incorrect. Fix issue with debugging arrays for TEST_FORCES not having indicies initialized. Other clean-up work.

sPuReMD: revert force tabulation changes. Temporarily work around out-of-bound memory accesses in cubic spline interpolation routes (for force tabulation) by expanding array sizes. Revisit at a later date.

sPuReMD: fix several more warning from Valgrind (out-of-bounds accesses in force tabulation lookup tables, sparse matrices with EE/ACKS2 with tabulation). Remove unused variables. Fixes to string handling. Other code clean-up.

sPuReMD: fix several issues with strings not being null-terminated. Fix use of uninitialized values reported by Valgrind (heap memory). Other general clean-ups.

sPuReMD: fix issue with list for debugging forces not having indices initialized (dDelta, dBO). Fix interaction counting in OpenMP code (greater than 1 thread). Other general clean-up.

sPuReMD: fix small differences in valence angle energy term with Fortran ReaxFF code. Other minor code updates.

sPuReMD: change log file column headers to be contigious strings of non-whitespace characters to facilitate post-processing.

Fix build issue where config.h.in is only generated once (for sPuReMD) => common file across all versions.

sPuReMD: fix issue with incorrect cutoff being applied for periodic neighbors in small simulation boxes (nonb_cut -> vlist_cut). Other clean-up and debug I/O updates.

sPuReMD: add ACKS2 geometry-dependent forces to tabulated computation of van der Waals and Coulomb terms. Replace double precision equality check with difference comparison.

sPuReMD: add missing header files when distributing (make dist). Fix OpenMP bug with ACKS2 geometry-dependent force contribution. Fix small OpenMP compilation issue. Update tools/run_sim.py to respect geo_format parameter and infer if not present for md_run_custom.

sPuReMD: default to vlist_buffer of 2.5 angstroms. Fixes to cutoffs used in far neighbor list generation.