source: pacpussensors/trunk/Vislab/lib3dv-1.2.0/lib3dv/eigen/doc/UsingIntelMKL.dox

Last change on this file was 136, checked in by ldecherf, 8 years ago

Doc

File size: 7.4 KB
Line 
1/*
2 Copyright (c) 2011, Intel Corporation. All rights reserved.
3 Copyright (C) 2011 Gael Guennebaud <gael.guennebaud@inria.fr>
4
5 Redistribution and use in source and binary forms, with or without modification,
6 are permitted provided that the following conditions are met:
7
8 * Redistributions of source code must retain the above copyright notice, this
9 list of conditions and the following disclaimer.
10 * Redistributions in binary form must reproduce the above copyright notice,
11 this list of conditions and the following disclaimer in the documentation
12 and/or other materials provided with the distribution.
13 * Neither the name of Intel Corporation nor the names of its contributors may
14 be used to endorse or promote products derived from this software without
15 specific prior written permission.
16
17 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
18 ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
19 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
20 DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
21 ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
22 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
23 LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
24 ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
26 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27
28 ********************************************************************************
29 * Content : Documentation on the use of Intel MKL through Eigen
30 ********************************************************************************
31*/
32
33namespace Eigen {
34
35/** \page TopicUsingIntelMKL Using Intel® Math Kernel Library from Eigen
36
37\section TopicUsingIntelMKL_Intro Eigen and Intel® Math Kernel Library (Intel® MKL)
38
39Since Eigen version 3.1 and later, users can benefit from built-in Intel MKL optimizations with an installed copy of Intel MKL 10.3 (or later).
40<a href="http://eigen.tuxfamily.org/Counter/redirect_to_mkl.php"> Intel MKL </a> provides highly optimized multi-threaded mathematical routines for x86-compatible architectures.
41Intel MKL is available on Linux, Mac and Windows for both Intel64 and IA32 architectures.
42
43\note
44Intel® MKL is a proprietary software and it is the responsibility of users to buy or register for community (free) Intel MKL licenses for their products. Moreover, the license of the user product has to allow linking to proprietary software that excludes any unmodified versions of the GPL.
45
46Using Intel MKL through Eigen is easy:
47-# define the \c EIGEN_USE_MKL_ALL macro before including any Eigen's header
48-# link your program to MKL libraries (see the <a href="http://software.intel.com/en-us/articles/intel-mkl-link-line-advisor/">MKL linking advisor</a>)
49-# on a 64bits system, you must use the LP64 interface (not the ILP64 one)
50
51When doing so, a number of Eigen's algorithms are silently substituted with calls to Intel MKL routines.
52These substitutions apply only for \b Dynamic \b or \b large enough objects with one of the following four standard scalar types: \c float, \c double, \c complex<float>, and \c complex<double>.
53Operations on other scalar types or mixing reals and complexes will continue to use the built-in algorithms.
54
55In addition you can coarsely select choose which parts will be substituted by defining one or multiple of the following macros:
56
57<table class="manual">
58<tr><td>\c EIGEN_USE_BLAS </td><td>Enables the use of external BLAS level 2 and 3 routines (currently works with Intel MKL only)</td></tr>
59<tr class="alt"><td>\c EIGEN_USE_LAPACKE </td><td>Enables the use of external Lapack routines via the <a href="http://www.netlib.org/lapack/lapacke.html">Intel Lapacke</a> C interface to Lapack (currently works with Intel MKL only)</td></tr>
60<tr><td>\c EIGEN_USE_LAPACKE_STRICT </td><td>Same as \c EIGEN_USE_LAPACKE but algorithm of lower robustness are disabled. This currently concerns only JacobiSVD which otherwise would be replaced by \c gesvd that is less robust than Jacobi rotations.</td></tr>
61<tr class="alt"><td>\c EIGEN_USE_MKL_VML </td><td>Enables the use of Intel VML (vector operations)</td></tr>
62<tr><td>\c EIGEN_USE_MKL_ALL </td><td>Defines \c EIGEN_USE_BLAS, \c EIGEN_USE_LAPACKE, and \c EIGEN_USE_MKL_VML </td></tr>
63</table>
64
65Finally, the PARDISO sparse solver shipped with Intel MKL can be used through the \ref PardisoLU, \ref PardisoLLT and \ref PardisoLDLT classes of the \ref PardisoSupport_Module.
66
67
68\section TopicUsingIntelMKL_SupportedFeatures List of supported features
69
70The breadth of Eigen functionality covered by Intel MKL is listed in the table below.
71<table class="manual">
72<tr><th>Functional domain</th><th>Code example</th><th>MKL routines</th></tr>
73<tr><td>Matrix-matrix operations \n \c EIGEN_USE_BLAS </td><td>\code
74m1*m2.transpose();
75m1.selfadjointView<Lower>()*m2;
76m1*m2.triangularView<Upper>();
77m1.selfadjointView<Lower>().rankUpdate(m2,1.0);
78\endcode</td><td>\code
79?gemm
80?symm/?hemm
81?trmm
82dsyrk/ssyrk
83\endcode</td></tr>
84<tr class="alt"><td>Matrix-vector operations \n \c EIGEN_USE_BLAS </td><td>\code
85m1.adjoint()*b;
86m1.selfadjointView<Lower>()*b;
87m1.triangularView<Upper>()*b;
88\endcode</td><td>\code
89?gemv
90?symv/?hemv
91?trmv
92\endcode</td></tr>
93<tr><td>LU decomposition \n \c EIGEN_USE_LAPACKE \n \c EIGEN_USE_LAPACKE_STRICT </td><td>\code
94v1 = m1.lu().solve(v2);
95\endcode</td><td>\code
96?getrf
97\endcode</td></tr>
98<tr class="alt"><td>Cholesky decomposition \n \c EIGEN_USE_LAPACKE \n \c EIGEN_USE_LAPACKE_STRICT </td><td>\code
99v1 = m2.selfadjointView<Upper>().llt().solve(v2);
100\endcode</td><td>\code
101?potrf
102\endcode</td></tr>
103<tr><td>QR decomposition \n \c EIGEN_USE_LAPACKE \n \c EIGEN_USE_LAPACKE_STRICT </td><td>\code
104m1.householderQr();
105m1.colPivHouseholderQr();
106\endcode</td><td>\code
107?geqrf
108?geqp3
109\endcode</td></tr>
110<tr class="alt"><td>Singular value decomposition \n \c EIGEN_USE_LAPACKE </td><td>\code
111JacobiSVD<MatrixXd> svd;
112svd.compute(m1, ComputeThinV);
113\endcode</td><td>\code
114?gesvd
115\endcode</td></tr>
116<tr><td>Eigen-value decompositions \n \c EIGEN_USE_LAPACKE \n \c EIGEN_USE_LAPACKE_STRICT </td><td>\code
117EigenSolver<MatrixXd> es(m1);
118ComplexEigenSolver<MatrixXcd> ces(m1);
119SelfAdjointEigenSolver<MatrixXd> saes(m1+m1.transpose());
120GeneralizedSelfAdjointEigenSolver<MatrixXd>
121 gsaes(m1+m1.transpose(),m2+m2.transpose());
122\endcode</td><td>\code
123?gees
124?gees
125?syev/?heev
126?syev/?heev,
127?potrf
128\endcode</td></tr>
129<tr class="alt"><td>Schur decomposition \n \c EIGEN_USE_LAPACKE \n \c EIGEN_USE_LAPACKE_STRICT </td><td>\code
130RealSchur<MatrixXd> schurR(m1);
131ComplexSchur<MatrixXcd> schurC(m1);
132\endcode</td><td>\code
133?gees
134\endcode</td></tr>
135<tr><td>Vector Math \n \c EIGEN_USE_MKL_VML </td><td>\code
136v2=v1.array().sin();
137v2=v1.array().asin();
138v2=v1.array().cos();
139v2=v1.array().acos();
140v2=v1.array().tan();
141v2=v1.array().exp();
142v2=v1.array().log();
143v2=v1.array().sqrt();
144v2=v1.array().square();
145v2=v1.array().pow(1.5);
146\endcode</td><td>\code
147v?Sin
148v?Asin
149v?Cos
150v?Acos
151v?Tan
152v?Exp
153v?Ln
154v?Sqrt
155v?Sqr
156v?Powx
157\endcode</td></tr>
158</table>
159In the examples, m1 and m2 are dense matrices and v1 and v2 are dense vectors.
160
161
162\section TopicUsingIntelMKL_Links Links
163- Intel MKL can be purchased and downloaded <a href="http://eigen.tuxfamily.org/Counter/redirect_to_mkl.php">here</a>.
164- Intel MKL is also bundled with <a href="http://software.intel.com/en-us/articles/intel-composer-xe/">Intel Composer XE</a>.
165
166
167*/
168
169}
Note: See TracBrowser for help on using the repository browser.