GenRTR Riemannian Optimization Package
GenRTR > tests > test_esgev.m

test_esgev

PURPOSE ^

TEST_ESGEV Test the RTRESGEV, IRTRESGEV and TMESGEV drivers

SYNOPSIS ^

function test_esgev(params,problem,p)

DESCRIPTION ^

 TEST_ESGEV   Test the RTRESGEV, IRTRESGEV and TMESGEV drivers

 This method tests the RTRESGEV, IRTRESGEV and TMESGEV drivers on a user-specified 
 generalized eigenvalue problem.
 If the user does not specify the problem, then the Matrix Market problem 
 BCSST_08 is used.
 
 test_esgev() uses default parameters.

 test_esgev(params) allows user-specification of parameters:
     params.x0        - initial iterate: x.U and x.V orthonormal
     params.rho_prime - trust-region acceptance parameters, default: .1
     params.useprec   - use a preconditioner (see rtresgev), default: 0
     params.testtm    - test tmesgev,   default: 1
     params.testrtr   - test rtresgev,  default: 1
     params.testirtr  - test irtresgev, default: 1

 test_esgev(params,problem) allows specification of the test matrix in a 
                        Matrix Market file format.
                        problem or problem{1} should specify filename for A
                        problem{2} should specify filename for B
                        default: problem = {'files/bcsstk08.mtx',
                                            'files/bcsstm08.mtx'}

 test_esgev(params,problem,p) allows user-specification of the number of 
                        basis vectors to be computed, default: 6

 See also rtr, irtr, irtresgev, rtresgev, tmesgev

CROSS-REFERENCE INFORMATION ^

This function calls:
  • irtresgev IRTRESGEV Compute extreme eigenvectors of a positive-definite Hermitian pencil
  • rtresgev RTRESGEV Compute extreme eigenvectors of a positive-definite Hermitian
  • tmesgev TMESGEV Compute extreme eigenvectors of a positive-definite Hermitian pencil
  • mmread
This function is called by:

Generated on Fri 11-Apr-2008 14:52:45 by m2html © 2003