topic Re:Lapack Diagonalization for Large Matrix in Intel® oneAPI Math Kernel Library

Lapack Diagonalization for Large Matrix

cd1054 — Thu, 19 Nov 2020 01:22:11 GMT

Hi,

I am working on finding the eigenvalues and eigenvectors of a matrix with dimension 52920. However, when I run the LAPACK routine, my program does not stop running after 16 hours. Do you know what might be wrong? I think there might be an infinite loop that is being entered, but it is difficult for me to track down the problem.

Here is part of my code for reference. Any help would be appreciated! Thank you.

void diagonalize(int D, double* ground_state, double* hamiltonian, double* z, double* w, int* ifail) {

cout << "dimension: " << D << endl;

int matrix_layout = LAPACK_ROW_MAJOR;

int dim = D;

int lda = D;

int ldz = D;

int il;

int iu;

long long int info;

int m;

double abstol = -1;

double vl = 0.0;

double vu = 100.0;

cout << "calling LAPACK..." << endl;

info = LAPACKE_dsyevx(matrix_layout,'V','A','L', dim, hamiltonian, lda, vl, vu, il, iu, abstol, &m, w, z, ldz,ifail);

Re:Lapack Diagonalization for Large Matrix

PrasanthD_intel — Thu, 19 Nov 2020 07:28:46 GMT

Hi,

There is a dedicated forum for mkl issues(https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/bd-p/oneapi-math-kernel-library), and we are redirecting issue to that forum.

Regards

Prasanth

Re: Re:Lapack Diagonalization for Large Matrix

cd1054 — Thu, 19 Nov 2020 17:44:11 GMT

Thank you!

It turns out that I was just being slightly impatient. The routine completed successfully after ~20 hours.

-Christina