HomeSort by relevance Sort by last modified time
    Searched full:bdcsvd (Results 1 - 15 of 15) sorted by null

  /external/eigen/test/
bdcsvd.cpp 23 #define SVD_DEFAULT(M) BDCSVD<M>
24 #define SVD_FOR_MIN_NORM(M) BDCSVD<M>
29 void bdcsvd(const MatrixType& a = MatrixType(), bool pickrandom = true) function
35 CALL_SUBTEST(( svd_test_all_computation_options<BDCSVD<MatrixType> >(m, false) ));
45 VERIFY_IS_APPROX(m.bdcSvd().singularValues(), RealVecType::Ones());
46 VERIFY_RAISES_ASSERT(m.bdcSvd().matrixU());
47 VERIFY_RAISES_ASSERT(m.bdcSvd().matrixV());
48 VERIFY_IS_APPROX(m.bdcSvd(ComputeFullU|ComputeFullV).solve(m), m);
56 BDCSVD<MatrixType> bdc_svd(m);
67 CALL_SUBTEST_3(( svd_verify_assert<BDCSVD<Matrix3f> >(Matrix3f()) ))
    [all...]
boostmultiprec.cpp 55 #include "bdcsvd.cpp"
199 CALL_SUBTEST_10(( bdcsvd(Mat(internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE), internal::random<int>(EIGEN_TEST_MAX_SIZE/4, EIGEN_TEST_MAX_SIZE/2))) ));
CMakeLists.txt 225 ei_add_test(bdcsvd)
  /external/eigen/failtest/
bdcsvd_int.cpp 13 BDCSVD<Matrix<SCALAR,Dynamic,Dynamic> > qr(Matrix<SCALAR,Dynamic,Dynamic>::Random(10,10));
  /external/eigen/Eigen/
SVD 24 * - BDCSVD implementing a recursive divide & conquer strategy on top of an upper-bidiagonalization which remains fast for large problems.
27 * - MatrixBase::bdcSvd()
38 #include "src/SVD/BDCSVD.h"
  /external/eigen/bench/
dense_solvers.cpp 57 BDCSVD<MatDyn> bdcsvd(A.rows(),A.cols());
72 BENCH(t_bdcsvd, tries, rep, bdcsvd.compute(A,svd_opt));
83 results["BDCSVD"][id] = t_bdcsvd.best();
98 labels.push_back("BDCSVD");
185 // cout << "BDCSVD (%) " << (results["BDCSVD"]/results["LLT"]).format(fmt) << "\n";
  /external/eigen/Eigen/src/SVD/
BDCSVD.h 31 template<typename _MatrixType> class BDCSVD;
36 struct traits<BDCSVD<_MatrixType> >
47 * \class BDCSVD
56 * For small matrice (<16), it is thus preferable to directly use JacobiSVD. For larger ones, BDCSVD is highly
67 class BDCSVD : public SVDBase<BDCSVD<_MatrixType> >
69 typedef SVDBase<BDCSVD> Base;
106 * perform decompositions via BDCSVD::compute(const MatrixType&).
108 BDCSVD() : m_algoswap(16), m_numIters(0)
116 * \sa BDCSVD()
    [all...]
SVDBase.h 45 * \sa class BDCSVD, class JacobiSVD
  /external/eigen/unsupported/bench/
bench_svd.cpp 48 BDCSVD<MatrixType> bdc_matrix(m);
79 BDCSVD<MatrixType> bdc_matrix(m, ComputeFullU|ComputeFullV);
  /external/tensorflow/tensorflow/core/kernels/
svd_op_impl.h 88 Eigen::BDCSVD<Matrix> svd(inputs[0], options);
  /external/eigen/doc/
DenseDecompositionBenchmark.dox 27 <tr class="alt"><td>BDCSVD</td><td>1.07 (x19.7)</td><td>21.83 (x51.5)</td><td>331.77 (x56.9)</td><td>18587.9 (x49.6)</td><td>110.53 (x16.3)</td><td>397.67 (x13.2)</td><td>2975 (x12.6)</td><td>48593.2 (x12.6)</td></tr>
35 + For large problem sizes, only the decomposition implementing a cache-friendly blocking strategy scale well. Those include LLT, PartialPivLU, HouseholderQR, and BDCSVD. This explain why for a 4k x 4k matrix, HouseholderQR is faster than LDLT. In the future, LDLT and ColPivHouseholderQR will also implement blocking strategies.
QuickReference.dox 21 <tr class="alt"><td>\link SVD_Module SVD \endlink</td><td>\code#include <Eigen/SVD>\endcode</td><td>SVD decompositions with least-squares solver (JacobiSVD, BDCSVD)</td></tr>
  /external/eigen/lapack/
svd.cpp 56 BDCSVD<PlainMatrixType> svd(mat,option);
  /external/eigen/Eigen/src/Core/
MatrixBase.h 375 inline BDCSVD<PlainObject> bdcSvd(unsigned int computationOptions = 0) const;
  /external/eigen/Eigen/src/Core/util/
ForwardDeclarations.h 259 template<typename MatrixType> class BDCSVD;

Completed in 235 milliseconds