Developer Reference for Intel® oneAPI Math Kernel Library for C

ID 766684
Date 7/13/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

Extended Eigensolver Output Details

Errors and warnings encountered during a run of the Extended Eigensolver routines are stored in an integer variable, info. If the value of the output info parameter is not 0, either an error or warning was encountered. The possible return values for the info parameter along with the error code descriptions are given in the following table.

Return Codes for info Parameter
info Classification Description

202

Error

Problem with size of the system n (n0)

201

Error

Problem with size of initial subspace m0 (m00 or m0>n)

200

Error

Problem with emin,emax (eminemax)

(100+i)

Error

Problem with i-th value of the input Extended Eigensolver parameter (fpm[i - 1]). Only the parameters in use are checked.

4

Warning

Successful return of only the computed subspace after call with fpm[13] = 1

3

Warning

Size of the subspace m0 is too small (m0<m)

2

Warning

No Convergence (number of iteration loops >fpm[3])

1

Warning

No eigenvalue found in the search interval. See remark below for further details.

0

Successful exit

 

-1

Error

Internal error for allocation memory.

-2

Error

Internal error of the inner system solver. Possible reasons: not enough memory for inner linear system solver or inconsistent input.

-3

Error

Internal error of the reduced eigenvalue solver

Possible cause: matrix B may not be positive definite. It can be checked by setting fpm[27] = 1 before calling an Extended Eigensolver routine, or by using LAPACK routines.

-4

Error

Matrix B is not positive definite.

-(100+i)

Error

Problem with the i-th argument of the Extended Eigensolver interface.

In some extreme cases the return value info=1 may indicate that the Extended Eigensolver routine has failed to find the eigenvalues in the search interval. This situation could arise if a very large search interval is used to locate a small and isolated cluster of eigenvalues (i.e. the dimension of the search interval is many orders of magnitude larger than the number of contour points. It is then either recommended to increase the number of contour points fpm[1] or simply rescale more appropriately the search interval. Rescaling means the initial problem of finding all eigenvalues the search interval [λmin,λmax] for the standard eigenvalue problem A x=λx is replaced with the problem of finding all eigenvalues in the search interval [λmin/ t, λmax/ t] for the standard eigenvalue problem (A/t) x=(λ/t) x where t is a scaling factor.