[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Getfem-commits] (no subject)
From: |
Tetsuo Koyama |
Subject: |
[Getfem-commits] (no subject) |
Date: |
Thu, 19 Sep 2019 06:36:12 -0400 (EDT) |
branch: fixmisspell
commit c1b2cc11ba015f703cf04f190c895e2f36b522a9
Author: Tetsuo Koyama <address@hidden>
Date: Thu Sep 19 19:35:56 2019 +0900
Fix typo in docs
---
doc/sphinx/source/gmm/blas.rst | 4 ++--
doc/sphinx/source/gmm/blas_interface.rst | 2 +-
doc/sphinx/source/gmm/denseqr.rst | 2 +-
doc/sphinx/source/gmm/iter.rst | 8 ++++----
doc/sphinx/source/gmm/matrix.rst | 2 +-
doc/sphinx/source/install/install_mac.rst | 2 +-
doc/sphinx/source/userdoc/model_plasticity_small_strain.rst | 2 +-
7 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/doc/sphinx/source/gmm/blas.rst b/doc/sphinx/source/gmm/blas.rst
index d4c9aa7..229f670 100644
--- a/doc/sphinx/source/gmm/blas.rst
+++ b/doc/sphinx/source/gmm/blas.rst
@@ -29,13 +29,13 @@ print to the standard output the vector ``V`` multiplied by
``10.0`` without cha
transposition
-------------
-``gmm::transposed(M)`` gives a possibily modifiable reference on the
transposed matrix of ``M``.
+``gmm::transposed(M)`` gives a possibility modifiable reference on the
transposed matrix of ``M``.
imaginary and real part
-----------------------
For a complex matrix ``M`` or a complex vector ``V``,
-``gmm::real_part(M)``, ``gmm::real_part(V)``, ``gmm::imag_part(M)`` or
``gmm::imag_part(V)`` give a possibily modifiable reference on the real or
imaginary part of the matrix or vector (for instance
``gmm::clear(gmm::imag_part(M))`` will set to zero the imaginary part of a
matrix ``M``). These functions cannot be applied to real matrices or vectors.
+``gmm::real_part(M)``, ``gmm::real_part(V)``, ``gmm::imag_part(M)`` or
``gmm::imag_part(V)`` give a possibility modifiable reference on the real or
imaginary part of the matrix or vector (for instance
``gmm::clear(gmm::imag_part(M))`` will set to zero the imaginary part of a
matrix ``M``). These functions cannot be applied to real matrices or vectors.
conjugate
---------
diff --git a/doc/sphinx/source/gmm/blas_interface.rst
b/doc/sphinx/source/gmm/blas_interface.rst
index bbdbdb7..7849100 100644
--- a/doc/sphinx/source/gmm/blas_interface.rst
+++ b/doc/sphinx/source/gmm/blas_interface.rst
@@ -10,7 +10,7 @@
Interface with BLAS, LAPACK or ATLAS
======================================
-For better performance on dense matrices, it is possible to interface some
operations of the type ``gmm::dense_matrix<T>`` with ``LAPACK``
(http://www.netlib.org/lapack/) or ``ATLAS``
(http://math-atlas.sourceforge.net/), for ``T = float, double,
std::complex<float> or std::complex<double>``. In fact, concerning ``ATLAS`` no
specific interface has been made untill now, so the fortran interface of
``ATLAS`` should be used.
+For better performance on dense matrices, it is possible to interface some
operations of the type ``gmm::dense_matrix<T>`` with ``LAPACK``
(http://www.netlib.org/lapack/) or ``ATLAS``
(http://math-atlas.sourceforge.net/), for ``T = float, double,
std::complex<float> or std::complex<double>``. In fact, concerning ``ATLAS`` no
specific interface has been made until now, so the fortran interface of
``ATLAS`` should be used.
to use this interface you have first to define ``GMM_USES_LAPACK`` before
including |gmm| \ files::
diff --git a/doc/sphinx/source/gmm/denseqr.rst
b/doc/sphinx/source/gmm/denseqr.rst
index 2e7f3d4..fa7538a 100644
--- a/doc/sphinx/source/gmm/denseqr.rst
+++ b/doc/sphinx/source/gmm/denseqr.rst
@@ -33,7 +33,7 @@ The following procedures are available in the file
``gmm/gmm\_dense\_qr.h`` for
-`Remark`: The computation of eigenvectors for non hermitian matrices is not
yet implemented. You can use for the moment the functions
``geev_interface_left`` and ``geev_interface_right`` from the LAPACK interface
(see ``gmm/gmm_lapack_interface.h``. These LAPACK functions compute right and
left eigen vectors.
+`Remark`: The computation of eigenvectors for non hermitian matrices is not
yet implemented. You can use for the moment the functions
``geev_interface_left`` and ``geev_interface_right`` from the LAPACK interface
(see ``gmm/gmm_lapack_interface.h``. These LAPACK functions compute right and
left eigenvectors.
The following function defined in the file ``gmm/gmm\_condition\_number.h``::
diff --git a/doc/sphinx/source/gmm/iter.rst b/doc/sphinx/source/gmm/iter.rst
index 290d1d5..97de0fb 100644
--- a/doc/sphinx/source/gmm/iter.rst
+++ b/doc/sphinx/source/gmm/iter.rst
@@ -11,7 +11,7 @@ Iterative solvers
-Most of the solvers provided in |gmm| come form ITL with slight modifications
(gmres has been optimized and adapted for complex matrices). Include the file
``gmm/gmm_iter_solvers.h`` to use them.
+Most of the solvers provided in |gmm| come frorm ITL with slight modifications
(gmres has been optimized and adapted for complex matrices). Include the file
``gmm/gmm_iter_solvers.h`` to use them.
iterations
----------
@@ -63,7 +63,7 @@ Here is the list of available linear solvers::
gmm::least_squares_cg(A, X, B, iter) // unpreconditionned least square CG.
-The solver ``gmm::constrained_cg(A, C, X, B, PS, PR, iter);`` solve a system
with linear constaints, ``C`` is a matrix which represents the constraints. But
it is still experimental.
+The solver ``gmm::constrained_cg(A, C, X, B, PS, PR, iter);`` solve a system
with linear constraints, ``C`` is a matrix which represents the constraints.
But it is still experimental.
(Version 1.7) The solver ``gmm::bfgs(F, GRAD, X, restart, iter)`` is a BFGS
quasi-Newton algorithm with a Wolfe line search for large scale problems. It
minimizes the function ``F`` without constraints, be given its gradient
``GRAD``. ``restart`` is the max number of stored update vectors.
@@ -104,7 +104,7 @@ The following preconditioners, to be used with linear
solvers, are available::
gmm::ilutp_precond<matrix_type> P(SM, k, threshold);
-Except ``ildltt\_precond``, all these precontionners come from ITL.
``ilut_precond`` has been optimized and simplified and ``cholesky_precond`` has
been corrected and transformed in an incomplete LDLT preconditioner for
stability reasons (similarly, we add ``choleskyt_precond`` which is in fact an
incomplete LDLT with threshold preconditioner). Of course, ``ildlt\_precond``
and ``ildltt_precond`` are designed for symmetric real or hermitian complex
matrices to be use principaly with cg.
+Except ``ildltt\_precond``, all these precontionners come from ITL.
``ilut_precond`` has been optimized and simplified and ``cholesky_precond`` has
been corrected and transformed in an incomplete LDLT preconditioner for
stability reasons (similarly, we add ``choleskyt_precond`` which is in fact an
incomplete LDLT with threshold preconditioner). Of course, ``ildlt\_precond``
and ``ildltt_precond`` are designed for symmetric real or hermitian complex
matrices to be use principally with cg.
Additive Schwarz method
-----------------------
@@ -119,7 +119,7 @@ For the moment, the method is not parallelized (this should
be done ...). The ca
The test program ``schwarz_additive.C`` is the directory ``tests`` of GetFEM++
is an example of the resolution with the additive Schwarz method of an
elastostatic problem with the use of coarse mesh to make a better
preconditioning (i.e. one of the sub-domains represents in fact a coarser mesh).
-In the case of multiple solves with the same linear system, it is possible to
store the preconditioners or the LU factorisations to save computation time.
+In the case of multiple solves with the same linear system, it is possible to
store the preconditioners or the LU factorizations to save computation time.
A (too) simple program in ``gmm/gmm_domain_decomp.h`` allows to build a
regular domain decomposition with a certain ratio of overlap. It directly
produces the vector of matrices ``vB`` for the additive Schwarz method.
diff --git a/doc/sphinx/source/gmm/matrix.rst b/doc/sphinx/source/gmm/matrix.rst
index 2ba67aa..aa3c4d0 100644
--- a/doc/sphinx/source/gmm/matrix.rst
+++ b/doc/sphinx/source/gmm/matrix.rst
@@ -114,7 +114,7 @@ The type ``gmm::csr_matrix<T>`` represents a compressed
sparse row matrix and ``
gmm::clean(M1, 1E-12);
gmm::copy(M1, M2);
-Matrices ``gmm::csr_matrix<T>`` and ``gmm::csc_matrix<T>`` have the advantage
to have a standard format (interfacable with Fortran code) and to have a
compact format (contiguous in memory). To be able to be compatible with Fortran
programs a second template parameter exists on these type, you can declare::
+Matrices ``gmm::csr_matrix<T>`` and ``gmm::csc_matrix<T>`` have the advantage
to have a standard format (interfaceable with Fortran code) and to have a
compact format (contiguous in memory). To be able to be compatible with Fortran
programs a second template parameter exists on these type, you can declare::
gmm::csc_matrix<double, 1> M1;
gmm::csr_matrix<double, 1> M2;
diff --git a/doc/sphinx/source/install/install_mac.rst
b/doc/sphinx/source/install/install_mac.rst
index 2d5ad1b..a7c47b1 100644
--- a/doc/sphinx/source/install/install_mac.rst
+++ b/doc/sphinx/source/install/install_mac.rst
@@ -159,7 +159,7 @@ Then, you will probably have to run
$ mex -setup
-To produce the correct ``mexopts.sh`` file in the ``.matlab/`` directory of
your home directory. If it still does not work, then you can try to modify the
``.matlab/mexopts.sh`` or replace it. Some ``mexopts.sh`` specially adpated to
macOS X/Xcode are available on the internet (See for instance here for
`MATLAB_R2015 <https://gist.github.com/varunagrawal/811e05ee4ca0f6a9952d>`_).
+To produce the correct ``mexopts.sh`` file in the ``.matlab/`` directory of
your home directory. If it still does not work, then you can try to modify the
``.matlab/mexopts.sh`` or replace it. Some ``mexopts.sh`` specially adapted to
macOS X/Xcode are available on the internet (See for instance here for
`MATLAB_R2015 <https://gist.github.com/varunagrawal/811e05ee4ca0f6a9952d>`_).
diff --git a/doc/sphinx/source/userdoc/model_plasticity_small_strain.rst
b/doc/sphinx/source/userdoc/model_plasticity_small_strain.rst
index 95c5dcc..e5308df 100644
--- a/doc/sphinx/source/userdoc/model_plasticity_small_strain.rst
+++ b/doc/sphinx/source/userdoc/model_plasticity_small_strain.rst
@@ -440,7 +440,7 @@ The yield condition still reads
.. math:: \delta \|B\| \le \sqrt{\Frac{2}{3}}(\sigma_{y0}+H_i \alpha_{n+1}).
-and for the elimination of the multiplier, :math:`\beta` has the same
expression as in the previous section adapting the value of :math:`\|B\|`. The
expressions of :math:`\bar{\zeta}_n` and :math:`\eta_n` have to be adpated
accoringly.
+and for the elimination of the multiplier, :math:`\beta` has the same
expression as in the previous section adapting the value of :math:`\|B\|`. The
expressions of :math:`\bar{\zeta}_n` and :math:`\eta_n` have to be adapted
accoringly.