svdin

来自「svd 算法代码 This directory contains instru」· 代码 · 共 42 行

#-----------------------------------
# Input file used for SVDRUN script
#-----------------------------------
#
# DESCRIPTION OF FIELDS:
#-----------------------------------------------------------------------------
#
# NUM --> Run number (1,2,3,...).
#
# CDE --> SVDPACKC code (i.e. las1, bls1, las2, bls2 etc.).
#
# FNM --> Sparse Matrix datafile (i.e. belladit, belladit, etc.).
#         (Path to datafiles specified in svdrun script)
#
# MXI --> Maximum number of iterations for method.
# 
# TRP --> Number of singular triplets desired.
#
# SUB --> Maximum subspace dimension.
#
# BSZ --> Initial blocksize size (if applicable).
#
# RED --> Residual norm reduction factor for initiating shifting. (tms1,tms2)
#
# RITZ --> Controls use of ritz shifting (0=no,1=yes). (tms1,tms2)
#
# ACC --> Accuracy (residual tolerance)
#
# VEC --> Compute singular vectors also? (TRUE/FALSE)
#-----------------------------------------------------------------------------
#
#  ID CDE     FNM      MXI  TRP  SUB  BSZ  RED   RITZ   ACC      VEC
#  -- ---     ---      ---  ---  ---  ---  ---   ----   ---      ---
   1  las1 belladit    50   10                          1.0E-6   TRUE
   2  las2 belladit    44   10                          1.0E-6   TRUE
   3  bls1 belladit    80   10   60    4                1.0E-6   TRUE
   4  bls2 belladit    40   10   20    5                1.0E-6   TRUE
   5  sis1 belladit   150   10   14                     1.0E-6   TRUE
   6  sis2 belladit    80   10   16                     1.0E-6   TRUE
   7  tms1 belladit    80   10   12        1.0E0   1    1.0E-6   TRUE
   8  tms2 belladit    80   10   12        1.0E0   1    1.0E-6   TRUE