Commits

Show all
Author Commit Message Labels Comments Date
Martin Albrecht
fixing warnings/errors reported by Microsoft Visual Studio
Tags
release-20091101
Martin Albrecht
another sizeof(size_t) != sizeof(word) bug
Martin Albrecht
fix bug which lead to wrong results on t2.math.washington.edu
Martin Albrecht
changing the soname version to 20091101 in preparation for new release
Martin Albrecht
merge
Martin Albrecht
fixed potential segmentaton fault in mzd_row_add_offset
Martin Albrecht
moving 'step 1.5' of LQUP MMPF to _mzd_lqup_submatrix because it caused confusion that the postprocessing is outside of that function.
Martin Albrecht
implemented timing experiment to calculate L1 and L2 cache size. This isn't working perfectly yet and thus it is only optional for now.
Martin Albrecht
whoops, forgot to check in configure.in
Martin Albrecht
don't check for the number of CPUs on configure. The macro is not cross platforms and we don't use it anyway (fixes #16)
Martin Albrecht
improve performance of mzd_transpose using Hacker's Delight bit-fiddling trick (closes: #15)
Martin Albrecht
use L2_CACHE_SIZE for PLUQ cutoff (experimental)
Martin Albrecht
copy submatrix to temporary when switching to MMPF
Martin Albrecht
added _pluq_mmpf back for debugging etc.
Martin Albrecht
don't apply permutation if todo rows == 0
Martin Albrecht
some performance improvements for sparse-ish matrices
Martin Albrecht
fixed a bug which escaped me for the last check in because I didnt check with cutoff=64
Martin Albrecht
improved performance for LQUP factorisation to roughly match that of PLUQ, still work to be done to improve upon PLUQ
Martin Albrecht
switching MMPF from PLUQ to LQUP and enabling it
Martin Albrecht
Added tag release-20090617 for changeset 46b89e01b348
Clement Pernet
Switch PLUQ -> LQUP Test suite passes Need optimizations
Martin Albrecht
made mzd_apply_p_right and mzd_apply_p_right_trans more efficient to decrease the penalty of column swaps.
Martin Albrecht
only swap at the end of the base case not during while finding the pivots. This allows a more performant search for pivots since we don't fill up the matrix with zero columns. This improves the 'sparse' case considerable. Joint work with Clement Pernet.
Martin Albrecht
implemented adding 3 and 4 rows in one step for PLUQ MMPF and adapted constants accordingly
Martin Albrecht
fix bug in mzd_is_zero() where small zero matrices wouldn't be reported as such
Martin Albrecht
switch back to using threads if any additional thread is available, don't require at least four
Martin Albrecht
added low leverl parallelisation ot process_rows2_pluq and added that the parallel sections in mzd_mul_mp_even() should use num_threads(4)
Martin Albrecht
fixing OpenMP doctest failures
Martin Albrecht
experiments with OpenMP
Martin Albrecht
fix compilation with --enable-openmp
  1. Prev
  2. Next