From 8d860282b4ee25679285c766a474ae4957bff5b4 Mon Sep 17 00:00:00 2001 From: bangerth Date: Fri, 6 Dec 2013 04:10:38 +0000 Subject: [PATCH] In the conversion to 64-bit indices, a lot of classes acquired a local typedef 'size_type'. In some cases, it was aliased to types::global_dof_index, but this may have been a mistake. One such example is FullMatrix, which definitely doesn't need 64-bit indices. Revert this and go back to 'unsigned int' for the local size_type type. git-svn-id: https://svn.dealii.org/trunk@31905 0785d39b-7218-0410-832d-ea1e28bc413d --- deal.II/doc/news/changes.h | 13 ++++ deal.II/include/deal.II/lac/full_matrix.h | 92 +++++++++++------------ deal.II/source/lac/full_matrix.inst.in | 14 ++-- 3 files changed, 66 insertions(+), 53 deletions(-) diff --git a/deal.II/doc/news/changes.h b/deal.II/doc/news/changes.h index 93c0ecd3a3..de3399598c 100644 --- a/deal.II/doc/news/changes.h +++ b/deal.II/doc/news/changes.h @@ -40,6 +40,19 @@ inconvenience this causes.

    +
  1. Changed: During the implementation of the 64-bit features for deal.II + 8.0, many linear algebra classes obtained a local typedef + size_type indicating the integer type that is used to index + into them. In some instances, this was accidentally set to + types::global_dof_index (which may be a 64-bit data type) + even in cases where this is clearly not going to work, for example for + FullMatrix::size_type, since we will not be able to store full matrix + objects of sizes for which a 32-bit index type is not sufficient. In + these cases, the typedef was reverted to just unsigned int. +
    + (Wolfgang Bangerth, 2013/12/04) +
    2. +
  2. Removed: With the switch of the testsuite to CMake, the old report_features and build test facilities are removed.
    diff --git a/deal.II/include/deal.II/lac/full_matrix.h b/deal.II/include/deal.II/lac/full_matrix.h index ad8a7fc09a..4829da893e 100644 --- a/deal.II/include/deal.II/lac/full_matrix.h +++ b/deal.II/include/deal.II/lac/full_matrix.h @@ -61,17 +61,17 @@ template class LAPACKFullMatrix; * * @author Guido Kanschat, Franz-Theo Suttmeier, Wolfgang Bangerth, 1993-2004 */ -template +template class FullMatrix : public Table<2,number> { public: /** - * Declare type of container size. + * A type of used to index into this container. Because we can not + * expect to store matrices bigger than what can be indexed by a regular + * unsigned integer, unsigned int is completely sufficient + * as an index type. */ - // A FullMatrix will never require to use unsigned long long int instead of - // unsigned but because a ConstraintMatrix can be a SparseMatrix or a - // FullMatrix, we need have the same interface. - typedef types::global_dof_index size_type; + typedef unsigned int size_type; /** * Type of matrix entries. In analogy to @@ -324,7 +324,7 @@ public: /** * Variable assignment operator. */ - template + template FullMatrix & operator = (const FullMatrix &); @@ -448,8 +448,8 @@ public: */ template void extract_submatrix_from (const MatrixType &matrix, - const std::vector &row_index_set, - const std::vector &column_index_set); + const std::vector &row_index_set, + const std::vector &column_index_set); /** * Copy the elements of the current matrix object into a specified @@ -487,7 +487,7 @@ public: * determined either by the size * of this or src. */ - template + template void fill (const FullMatrix &src, const size_type dst_offset_i = 0, const size_type dst_offset_j = 0, @@ -499,7 +499,7 @@ public: * Make function of base class * available. */ - template + template void fill (const number2 *); /** @@ -520,7 +520,7 @@ public: * possible to duplicate rows or * columns by this method. */ - template + template void fill_permutation (const FullMatrix &src, const std::vector &p_rows, const std::vector &p_cols); @@ -612,7 +612,7 @@ public: * complex-valued, but not mixed, for * this function to make sense. */ - template + template number2 matrix_norm_square (const Vector &v) const; /** @@ -630,7 +630,7 @@ public: * complex-valued, but not mixed, for * this function to make sense. */ - template + template number2 matrix_scalar_product (const Vector &u, const Vector &v) const; @@ -825,7 +825,7 @@ public: * convertible to the data type * of this matrix. */ - template + template void add (const number a, const FullMatrix &A); @@ -842,7 +842,7 @@ public: * convertible to the data type * of this matrix. */ - template + template void add (const number a, const FullMatrix &A, const number b, @@ -861,7 +861,7 @@ public: * is convertible to the data * type of this matrix. */ - template + template void add (const number a, const FullMatrix &A, const number b, @@ -886,7 +886,7 @@ public: * size of this or src * and the given offsets. */ - template + template void add (const FullMatrix &src, const number factor, const size_type dst_offset_i = 0, @@ -901,7 +901,7 @@ public: * * A += s BT */ - template + template void Tadd (const number s, const FullMatrix &B); @@ -927,7 +927,7 @@ public: * of this or * src. */ - template + template void Tadd (const FullMatrix &src, const number factor, const size_type dst_offset_i = 0, @@ -959,10 +959,10 @@ public: * and are indeed ignored in the * implementation. */ - template + template void add (const size_type row, - const size_type n_cols, - const size_type *col_indices, + const unsigned int n_cols, + const index_type *col_indices, const number2 *values, const bool elide_zero_values = true, const bool col_indices_are_sorted = false); @@ -1030,7 +1030,7 @@ public: * Assignment *this = * a*A. */ - template + template void equ (const number a, const FullMatrix &A); @@ -1038,7 +1038,7 @@ public: * Assignment *this = a*A + * b*B. */ - template + template void equ (const number a, const FullMatrix &A, const number b, @@ -1048,7 +1048,7 @@ public: * Assignment *this = a*A + * b*B + c*C. */ - template + template void equ (const number a, const FullMatrix &A, const number b, @@ -1176,7 +1176,7 @@ public: * usually results in considerable * performance gains. */ - template + template void mmult (FullMatrix &C, const FullMatrix &B, const bool adding=false) const; @@ -1208,7 +1208,7 @@ public: * BLAS usually results in considerable * performance gains. */ - template + template void Tmmult (FullMatrix &C, const FullMatrix &B, const bool adding=false) const; @@ -1240,7 +1240,7 @@ public: * usually results in considerable * performance gains. */ - template + template void mTmult (FullMatrix &C, const FullMatrix &B, const bool adding=false) const; @@ -1273,7 +1273,7 @@ public: * BLAS usually results in considerable * performance gains. */ - template + template void TmTmult (FullMatrix &C, const FullMatrix &B, const bool adding=false) const; @@ -1322,7 +1322,7 @@ public: * Source and destination must * not be the same vector. */ - template + template void vmult (Vector &w, const Vector &v, const bool adding=false) const; @@ -1334,7 +1334,7 @@ public: * Source and destination must * not be the same vector. */ - template + template void vmult_add (Vector &w, const Vector &v) const; @@ -1357,7 +1357,7 @@ public: * Source and destination must * not be the same vector. */ - template + template void Tvmult (Vector &w, const Vector &v, const bool adding=false) const; @@ -1370,7 +1370,7 @@ public: * Source and destination must * not be the same vector. */ - template + template void Tvmult_add (Vector &w, const Vector &v) const; @@ -1398,7 +1398,7 @@ public: * dst must not be the same * vector. */ - template + template number residual (Vector &dst, const Vector &x, const Vector &b) const; @@ -1420,7 +1420,7 @@ public: * same object for @p dst and @p * src. */ - template + template void forward (Vector &dst, const Vector &src) const; @@ -1434,7 +1434,7 @@ public: * same object for @p dst and @p * src. */ - template + template void backward (Vector &dst, const Vector &src) const; @@ -1576,8 +1576,8 @@ template inline void FullMatrix::extract_submatrix_from (const MatrixType &matrix, - const std::vector &row_index_set, - const std::vector &column_index_set) + const std::vector &row_index_set, + const std::vector &column_index_set) { AssertDimension(row_index_set.size(), this->n_rows()); AssertDimension(column_index_set.size(), this->n_cols()); @@ -1627,7 +1627,7 @@ FullMatrix::set (const size_type i, template -template +template void FullMatrix::vmult_add (Vector &w, const Vector &v) const @@ -1637,7 +1637,7 @@ FullMatrix::vmult_add (Vector &w, template -template +template void FullMatrix::Tvmult_add (Vector &w, const Vector &v) const @@ -1833,13 +1833,13 @@ FullMatrix::add (const size_type r, const size_type c, const number v) template -template +template inline void -FullMatrix::add (const size_type row, - const size_type n_cols, - const size_type *col_indices, - const number2 *values, +FullMatrix::add (const size_type row, + const unsigned int n_cols, + const index_type *col_indices, + const number2 *values, const bool, const bool) { diff --git a/deal.II/source/lac/full_matrix.inst.in b/deal.II/source/lac/full_matrix.inst.in index f902a72697..7484bbd483 100644 --- a/deal.II/source/lac/full_matrix.inst.in +++ b/deal.II/source/lac/full_matrix.inst.in @@ -51,7 +51,7 @@ for (S1, S2 : REAL_SCALARS) template void FullMatrix::fill ( - const FullMatrix&, types::global_dof_index, types::global_dof_index, types::global_dof_index, types::global_dof_index); + const FullMatrix&, size_type, size_type, size_type, size_type); template void FullMatrix::add (const S1, const FullMatrix&); template @@ -63,12 +63,12 @@ for (S1, S2 : REAL_SCALARS) const S1, const FullMatrix&); template void FullMatrix::add ( - const FullMatrix&, S1, types::global_dof_index, types::global_dof_index, types::global_dof_index, types::global_dof_index); + const FullMatrix&, S1, size_type, size_type, size_type, size_type); template void FullMatrix::Tadd (const S1, const FullMatrix&); template void FullMatrix::Tadd ( - const FullMatrix&, S1, types::global_dof_index, types::global_dof_index, types::global_dof_index, types::global_dof_index); + const FullMatrix&, S1, size_type, size_type, size_type, size_type); template void FullMatrix::equ (const S1, const FullMatrix&); template @@ -159,7 +159,7 @@ for (S1, S2 : COMPLEX_SCALARS) { template void FullMatrix::fill ( - const FullMatrix&, types::global_dof_index, types::global_dof_index, types::global_dof_index, types::global_dof_index); + const FullMatrix&, size_type, size_type, size_type, size_type); template void FullMatrix::add (const S1, const FullMatrix&); template @@ -171,13 +171,13 @@ for (S1, S2 : COMPLEX_SCALARS) const S1, const FullMatrix&); template void FullMatrix::add ( - const FullMatrix&, S1, types::global_dof_index, types::global_dof_index, types::global_dof_index, types::global_dof_index); + const FullMatrix&, S1, size_type, size_type, size_type, size_type); template void FullMatrix::Tadd (const S1, const FullMatrix&); template void FullMatrix::Tadd ( - const FullMatrix&, S1, types::global_dof_index, types::global_dof_index, - types::global_dof_index, types::global_dof_index); + const FullMatrix&, S1, size_type, size_type, + size_type, size_type); template void FullMatrix::equ (const S1, const FullMatrix&); template -- 2.39.5