topic Re:mkl_sparse_d_add vs mkl_dcsradd for repeatedly adding matrices of identical structure in Intel® oneAPI Math Kernel Library

mkl_sparse_d_add vs mkl_dcsradd for repeatedly adding matrices of identical structure

Jens_E_ — Fri, 30 Jul 2021 09:00:08 GMT

Hi,

The deprecated mkl_?crsadd routines has a 'request' parameter which allows for some precomputation/analysis of matrix structure when matrices of the same structure are to be added repeatedly.

Is there a similiar option within the new OneApi methods (mkl_sparse_?_add or related)?

Jens

Re:mkl_sparse_d_add vs mkl_dcsradd for repeatedly adding matrices of identical structure

Gennady_F_Intel — Fri, 30 Jul 2021 11:27:01 GMT

Jens,

that's not possible with this Inspector - Executing API. You might submit the official FR to the Intel Online Service Center channel regarding this topic.

Re:mkl_sparse_d_add vs mkl_dcsradd for repeatedly adding matrices of identical structure

Gennady_F_Intel — Sat, 31 Jul 2021 03:38:52 GMT

you might take a look at the

mkl_sparse_sp2m routine from IE set of Sparse BLAS :

https://software.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/blas-and-sparse-blas-routines/inspector-executor-sparse-blas-routines/inspector-executor-sparse-blas-execution-routines/mkl-sparse-sp2m.html

Re: mkl_sparse_d_add vs mkl_dcsradd for repeatedly adding matrices of identical structure

Jens_E_ — Mon, 02 Aug 2021 07:04:50 GMT

Thanks!

Re:mkl_sparse_d_add vs mkl_dcsradd for repeatedly adding matrices of identical structure

Gennady_F_Intel — Tue, 03 Aug 2021 12:25:19 GMT

The issue is closing and we will no longer respond to this thread. If you require additional assistance from Intel, please start a new thread. Any further interaction in this thread will be considered community only.