Macaulay2 » Documentation
Packages » Macaulay2Doc » matrices » SVD » SVD(...,DivideConquer=>...)
next | previous | forward | backward | up | index | toc

SVD(...,DivideConquer=>...) -- whether to use the LAPACK divide and conquer SVD algorithm

Description

For large matrices, this algorithm is often much faster.
i1 : M = random(RR^200, RR^200);

                200         200
o1 : Matrix RR      <-- RR
              53          53
i2 : time SVD(M);
 -- used 0.0620269s (cpu); 0.0163723s (thread); 0s (gc)
i3 : time SVD(M, DivideConquer=>true);
 -- used 0.543806s (cpu); 0.190059s (thread); 0s (gc)

Functions with optional argument named DivideConquer:

Further information

  • Default value: true
  • Function: SVD -- singular value decomposition of a matrix
  • Option key: DivideConquer -- an optional argument

The source of this document is in Macaulay2Doc/functions/SVD-doc.m2:66:0.