topic Re: dsyev Performance in Intel® oneAPI Math Kernel Library

dsyev Performance

Deleted_U_Intel — Thu, 15 Jun 2006 03:32:00 GMT

Dear,

Is it correct that the dsyev function call is about 6 times slower than Jacobi from numerical recipes (11.1)for 3x3matriches ?

Is it possible to use the dsyev call with two dimensional arrays?

All feedback is highly appreciated!

Thanks,

Bart

Chao_Y_Intel — Tue, 20 Jun 2006 13:48:28 GMT

Hi Bart,

Intel MKL is high optimized for large matrix operation. 3x3 matrixes are very small matrix for Intel MKL. MKL are not high optimized for them

Regards,
Chao

Chao_Y_Intel — Tue, 20 Jun 2006 13:48:44 GMT

Hi Bart,

Intel MKL is high optimized for large matrix operation. 3x3 matrixes are very small matrix for Intel MKL. MKLis not high optimized for them

Regards,
Chao