77 lines
4.8 KiB
Plaintext
77 lines
4.8 KiB
Plaintext
LA: linear algebra C++ interface library
|
|
|
|
Copyright (C) 2008- Jiri Pittner <jiri.pittner@jh-inst.cas.cz> or <jiri@pittnerovi.com>
|
|
complex versions written by Roman Curik <roman.curik@jh-inst.cas.cz>
|
|
cuda interface contributed by Miroslav Sulc <miroslav.sulc@jh-inst.cas.cz>
|
|
|
|
This program is free software: you can redistribute it and/or modify
|
|
it under the terms of the GNU General Public License as published by
|
|
the Free Software Foundation, either version 3 of the License, or
|
|
(at your option) any later version.
|
|
|
|
This program is distributed in the hope that it will be useful,
|
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
GNU General Public License for more details.
|
|
|
|
You should have received a copy of the GNU General Public License
|
|
along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
--------------------------------------------------------------------------------------------------
|
|
This software provides a C++ vector and matrix class with an interface
|
|
to BLAS and ATLAS linear algebra libraries and a few additional features.
|
|
Templates are employed in order to achieve generic applicability of the algorithms.
|
|
In particular, iterative methods suitable for sparse matrices (Davidson diagonalization, linear solvers, matrix exponential)
|
|
can be applied to your custom matrix class, which does not need to contain any explicit
|
|
storage of the matrix (only matrix times vector operation has to be provided).
|
|
(In quantum chemistry, a particular application of this technique is called direct-CI.)
|
|
The library implements reference counting to avoid overhead when copying matrices and passing them by value;
|
|
on the other hand it does NOT optimize matrix expressions, you should yourself call gemv, axpy etc. instead of using operator* etc. where appropriate.
|
|
|
|
vec.h, vec.cc : vector class
|
|
mat.h, mat.cc : general matrix class
|
|
smat.h, smat.cc: symmetric matrix class
|
|
vecmat3.h, vecmat3.cc : simplified class for 3-dimensional entities
|
|
quaternion.h, quaternion.cc : quaternions and 3-dim rotations
|
|
sparsemat.h sparsemat.cc: sparse matrix class
|
|
sparsesmat.h sparsesmat.cc: sparse symmetric matrix class
|
|
csrmat.h, csrmat.cc: unfinished work on compressed row format sparse matrices
|
|
nonclass.h nonclass.cc: some methods not belonging to a class - linear solver, diagonalization
|
|
conjgrad.h: conjugate gradient sparse linear solver
|
|
gmres.h: generalized minimal residual sparse linear solver
|
|
matexp.h: matrix exponential (suitable both for dense and sparse matrices - cf. exptimes routine)
|
|
qsort.h: generic quick-sort template
|
|
bitvector.h: bit vector class
|
|
bisection.h: generic bisection search
|
|
diis.h: DIIS convergence acceleration
|
|
davidson.h: Davidson (modified Lanczos) sparse matrix diagonalization
|
|
fourindex.h, fourindex.cc: class for four-index quantities, in particular two-electron integrals in quantum chemistry, allows transparent access to externally stored integrals
|
|
auxstorage.h: class facilitating simple file IO for the vectors and matrices
|
|
permutation.h, qsort.h : permutations and sorting
|
|
|
|
|
|
INSTALLATION:
|
|
0. autoreconf --install if ./configure does not exist
|
|
1. ./configure with DEBUG, OPTIMIZE and MATPTR options, export CXXFLAGS=-I<path> and export LDFLAGS=-L<path> (and possibly added -latlas -lcblas)
|
|
pointing to your preferred BLAS/LAPACK, your preferred --prefix and --enable/--disable options for optimization and debugging
|
|
2. make
|
|
3. make check (to see that the programs t and test compile and link without errors)
|
|
4. make install
|
|
|
|
USING:
|
|
1. #include "la.h"
|
|
2. using namespace LA;
|
|
3. look at test.cc and t.cc for examples of use
|
|
4. remember to always call .copyonwrite() before you start writing to individual matrix elements explicitly
|
|
(when modifying a matrix via provided methods or operators other than () and [], copy on write is performed transparently)
|
|
5. Note that the library does not attempt to do any optimizations of the evaluation of matrix expressions (by template metaprogramming and postponed evaluation techniques), so instead of inefficient "z=A*2*x + y*1.5 " call gemv and axpy methods. On the other hand it makes shallow copies, so you can pass matrix by value to a subroutine and you are allowed to write x=A*x, as a temporary will be created.
|
|
6. you may compile it with DEBUG defined to perform index range and other checks
|
|
|
|
NOTE:
|
|
We named the classes NRMat, NRVec etc. since we started to use them together with the Numerical Recipes library. They can be used with Numerical Recipes in C++, however, they are completely independent of this library.
|
|
|
|
This is a beta release. The library has already been in use, but not every function/method has been tested.
|
|
There is no documentation except the source itself, since I originaly did not intend to distribute it.
|
|
I will appreciate your bugfixes, exhancements, suggestions etc.
|
|
|
|
Jiri Pittner
|