handling sparsity of matrix in intel CMKL 8.1

handling sparsity of matrix in intel CMKL 8.1

sonuswaika's picture

Hi,

I want to calculate all the eigenvalues of a real, symmetric, sparse matrix of the order of 30,000 X 30,000.
Since, the order of the matrix is so huge, I would like to store only the non-zero values of the matrix to cut down on memory. I have been using scalapack earlier (pdsyev subroutine) for calculating the eigenvalues but unfortunately for that i have to store even the zeroes.
can anyone please suggest me that is there any provision in intel cmkl to take care of the sparsity of the matrix or can there be any other method so that i don't fall short of memory.


thanks
Sonu

1 post / 0 new
For more complete information about compiler optimizations, see our Optimization Notice.