Return to the 'sparsersb' package
Summary of important user-visible changes for releases of the sparsersb package =============================================================================== sparsersb-1.0.9 Release Date: 2021-09-21 =============================================================================== ** Intended to be used with librsb-1.2.0.9 or later. ** - documented differences with sparse in multiplying sparsersb by sparsersb ** - documented differences with sparse in summing sparsersb by sparsersb ** - documented "symmetric" and "hermitian" specifiers ** - documented "symmetric" and "hermitian" (i,j) access ** - documented "symmetric" and "hermitian" (:,:) access ** - fix: conversion from symmetric complex sparsersb won't give hermitian ** - fix: conversion from diagonal complex sparse shan't give hermitian ** - fix: compatibility with Octave-6 in sparsersbtg.m ** - fix: avoid test requiring >1 threads failing on OMP_NUM_THREADS=1 ** - fix: avoid possible segfault crash (heap-use-after-free) ** - better message when trying to invert matrix with zeroes on diagonal ** - stricter error propagation when in `make tests' ** - more error verbosity =============================================================================== sparsersb-1.0.8 Release Date: 2020-08-06 =============================================================================== ** Intended to be used with librsb-1.2.0.9 or later. ** - fixed '*' giving bad results between real sparsersb and complex scalar ** - fixed '+' giving bad sum between real sparsersb and complex sparsersb ** - improved documentation, demos and tests ** - fix: hermitian sparsersb matrices were being marked with 'U', not 'H' ** - fix: avoid duplicate vector re-zeroing (so, a bit faster) =============================================================================== sparsersb-1.0.7 Release Date: 2020-01-07 =============================================================================== ** - fixed '*' operator: 2*sparsersb(...) now returns a sparsersb, not sparse ** - fixed '*' operation when one operand is real sparsersb and other complex ** - avoids internal octave API deprecated in 4.4 ** - builds with 4.4 till pre-release 6.0 ** - `make tests' shall fail on sparsersb not matching octave's sparse ** - implemented elemental transposition by complex: sparsersb.^complex ** - expand symmetry when converting sparsersb double to sparse ** - expand hermitian when converting sparsersb complex to sparse ** - 1-D indexing read access is meant to be like in sparse ** - (:), (:,:), (i,:), (:,j) access support implemented (but not efficient) ** - internal bug fixes ** - more test cases (for 'test librsb') =============================================================================== sparsersb-1.0.6 Release Date: 2018-06-15 =============================================================================== ** Intended to be used with the latest librsb-1.2.0. ** - fix of typo breaking builds (bug # #50764) =============================================================================== sparsersb-1.0.5 Release Date: 2017-03-29 =============================================================================== ** Intended to be used with the latest librsb-1.2.0. ** - tests won't use quit(): might emit exception according to bug #49271 =============================================================================== sparsersb-1.0.4 Release Date: 2017-03-25 =============================================================================== ** Intended to be used with the latest librsb-1.2.0. ** - test script now uses ilu() instead of obsolete luinc() =============================================================================== sparsersb-1.0.3 Release Date: 2017-03-24 =============================================================================== ** Intended to be used with the latest librsb-1.2.0. ** - "symmetric" RSB representation supported ** - improved documentation (including a few typos) ** - improved demos (demo sparsersb) ** - improved tests (test sparsersb) ** - improved error messages =============================================================================== sparsersb-1.0.2 Release Date: 2016-10-03 =============================================================================== ** Intended to be used with the latest librsb-1.2.0. ** - builds even if octave built with --enable-64, but limited to matrices ** which would fit when using a normal setup. =============================================================================== sparsersb-1.0.1 Release Date: 2016-08-01 =============================================================================== ** Intended to be used with the latest librsb-1.2.0. ** Changed sparsersb's `configure --help': ** - options to build librsb from a tarball (via configure or LIBRSB_TARBALL) ** - options to use librsb-config ** - options to override librsb-config ** - you can override the default C++11 flag =============================================================================== sparsersb-1.0.0 Release Date: 2015-05-31 =============================================================================== ** First Packaged Release. Intended to work with librsb-1.2.
Package: sparsersb