LICENSE

DIRECT PARTIAL SVD Solver using xscqr, xGEBRD, and dstevx2
ddpsvd.f   zdpsvd.f

Computing orthonormal bases for Sakurai-Sugiura method using xscqr, xGEBRD, ddqds, and xoqdsq
dorth.f   zunth.f

SHIFTED CHOLESKY QR DECOMPOSITION for a tall and skinny matrix
dscqr.f   sscqr.f   zscqr.f   cscqr.f   dsclq.f   ssclq.f   zsclq.f   csclq.f

A new implementation of DQDS ALGORITHM WITH NEW SHIFT STRATEGIES
ddqds.f   sdqds.f

A new implementation of MDLVS ALGORITHM WITH NEW SHIFT STRATEGIES
dmdlvs.f

PARALLEL BISECTION METHOD for a symmetric tridiagonal matrix
dlaebz2.f   dstebz2.f   dstevx2.f

OpenMP-based PARALLEL BLOCKED INVERSE ITERATION ALGORITHM with DGEMM-based blocked classical Gram-Schmidt*2 reorthogonalization for a symmetric tridiagonal matrix
dstein3.f

PARALLEL BISECTION METHOD and INVERSE ITERATION ALGORITHM for a symmetric band matrix
BBiInv.tgz

A new implementation of ORTHOGONAL QD ALGORITHM WITH NEW SHIFT STRATEGIES
This code can compute singular values, left singular vectors and right singular vectors of a lower/upper bidiagonal matrix.
doqds.f   soqds.f   zoqds.f   coqds.f
This code can compute the desired number of singular pairs (pairs of singular values and right singular vectors) of the lower bidiagonal matrix. The desired number means the number of singular values counted from the smaller singular value.
doqdsq.f   soqdsq.f   zoqdsq.f   coqdsq.f
This code can compute only singular values of a lower/upper bidiagonal matrix.
doqdsv.f   doqdsv.f
Subroutines for ORTHOGONAL QD ALGORITHM WITH NEW SHIFT STRATEGIES
dlartg6.f   dlartg7.f   dfma0.c   dlas2u.f   drot2.f   slartg6.f   slartg7.f   sfma0.c   slas2u.f   srot2.f   csrot2.f   zdrot2.f  

Augmented implicitly restarted Lanczos bidiagonalization methods
DTRGKL5.tgz   STRGKL5.tgz   ZTRGKL5.tgz   CTRGKL5.tgz

Thick Restart Lanczos method
DTRLAN3.tgz   STRLAN3.tgz   ZTRLAN3.tgz   CTRLAN3.tgz

Singular Value Decomposition and Principal Component Analysis using ARPACK
debug.h   dsvd_file.f90   dpca_file.f90   svd_make_test_file.c   ARmake.inc for Windows

RANDOMIZED SVD
RANDOMIZED_SVD.tgz

JACOBI EVD for an upper triangular part of a symmetric matrix
djacobievd.f90   sjacobievd.f90   zjacobievd.f90   cjacobievd.f90

TWO SIDED JACOBI SVD for an upper triangular matrix
djacobisvd.f90   sjacobisvd.f90   zjacobisvd.f90   cjacobisvd.f90

ONE SIDED JACOBI SVD for a general matrix
donesidejacobi.f90   sonesidejacobi.f90   zonesidejacobi.f90   conesidejacobi.f90

Bidiagonalization for a complex upper triangular matrix
zbidiag.f90   cbidiag.f90

LAPROGNC is developed by the following researchers.

• Kyoto University: Yoshimasa NAKAMURA, Kinji KIMURA, Takumi YAMASHITA, Hiroyuki ISHIGAMI, Hiroki TANAKA, Yuki FUJII, Sho ARAKI, Yuya ISHIDA, Masayuki OSAWA, Masana AOKI
• Nara Women's University: Masami TAKATA, Miho CHIYONOBU
• Kyushu University: Katsuki FUJISAWACHayato WAKI
• Kyoto Prefectural University: Masashi IWASAKI
• The University of Electro-Communications: Yusaku YAMAMOTO
• University of Fukui: Hiroki WATANABE, Takahiro MIYAMAE, Riku HARANO, NOR ALIAH BINTI BAHARDIN, Rika TANAKA, SITI MAHIRAH BINTI MOHD NOR, Shota NISHIYAMA, Noriyuki TAKAMURA, Toshihiro NISHIKAWA, Seiya FUJITA, Jun HARAYAMA, Muhammad Aiman Bin Syed Shamsudin Syed

For any quetion, contact us via "kimura.kinji.7z -*- kyoto-u.jp" (please change "-*-" into "@")