From: wolf Date: Wed, 17 May 2000 13:53:40 +0000 (+0000) Subject: apply_boundary_values. jacobians X-Git-Url: https://gitweb.dealii.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=f634134226076fd4f4e0bf551c1623f1aa211513;p=dealii-svn.git apply_boundary_values. jacobians git-svn-id: https://svn.dealii.org/trunk@2876 0785d39b-7218-0410-832d-ea1e28bc413d --- diff --git a/deal.II/doc/news/2000/c-3-0.html b/deal.II/doc/news/2000/c-3-0.html index d647dffbdd..46b8ecf079 100644 --- a/deal.II/doc/news/2000/c-3-0.html +++ b/deal.II/doc/news/2000/c-3-0.html @@ -304,6 +304,40 @@
(WB 2000/05/16)

+ +
  • + Changed: The computation of the Jacobian matrices in the + FEValues class is now done more + efficiently. The speedup is in the range of a factor of 40 for + 3D. +
    + (John Burnell, WB 2000/05/16) +

    + +
  • + Changed: the + MatrixTools::apply_boundary_values + now uses a much faster algorithm when using on matrices with + symmetric sparsity patterns. On the other hand, it does no more + eliminate whole rows when a matrix has a non-symmetric sparsity + pattern, or if the user (through a new flag) tells the function + that this is not necessary, for example if the matrix itself is + non-symmetric. +

    + +

    + For symmetric sparsity patterns, the algorithm now eliminates + each boundary value in O(m*log(m)) steps instead of + O(N*log(m)), where N=number of rows of the matrix, and m=number + of entries per row. Note that m is roughly constant, + irrespective of N, so the old algorithm became slower with + finer grids, while the new one is O(1) for each boundary degree + of freedom. +

    + +

    + (John Burnell, WB 2000/05/17) +