topic MKL Diagonal SpMV in Intel® oneAPI Math Kernel Library

MKL Diagonal SpMV

Mohammad_A_ — Mon, 25 Jul 2016 18:02:17 GMT

Hi,

Is mkl_?diagemv multi-threaded ?

Thanks,

Mohammad Almasri

Alexander_K_Intel2 — Tue, 26 Jul 2016 07:12:31 GMT

Hi Mohammad,

Current implementation of matrix vector multiplication for diag format implemented sequentially for 1 array and parallel for several arrays.

Thanks,

Alex

Mohammad_A_ — Tue, 26 Jul 2016 15:28:54 GMT

Hi,

What I understand that this routine compute the SpMV for only one sparse matrix and one input vector, What do you mean by several arrays ?

Thanks,

Mohammad Almasri