Showing results for

- Intel Community
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library & Intel® Math Kernel Library
- SPARSE_STATUS_NOT_SUPPORTED exit status in mkl_sparse_sp2m with symmetric sparse matrices

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page

Highlighted
##

yera__rolando

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

08-18-2019
08:41 PM

36 Views

SPARSE_STATUS_NOT_SUPPORTED exit status in mkl_sparse_sp2m with symmetric sparse matrices

Hi

I'm trying to multiply to symmetric sparse matrices using the mkl_sparse_sp2m function, the exit status is SPARSE_STATUS_NOT_SUPPORTED and i don't understand why, here goes the reproducer. I'm using mkl 2019 update 4 in debian.

4 Replies

Highlighted
##

Gennady_F_Intel

Moderator

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

08-19-2019
03:04 AM

36 Views

That's mean the current implementation doesn't cover this set of input options.

Highlighted
##

Kirill_V_Intel

Employee

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

08-19-2019
02:09 PM

36 Views

Hello,

Gennady is right, this case is not supported currently in mkl_sparse_sp2m.

Though this case is supported by mkl_sparse_spmm as far as I see. So, is there a particular reason why you want to call sp2m here?

Best,

Kirill

Highlighted
##

yera__rolando

Beginner

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

08-20-2019
07:48 AM

36 Views

Hello:

Yes this particular case is supported by mkl_sparse_spmm, the reason why I use mkl_sparse_sp2m is because in my real application I wanted to store only the symmetric part of the matrix on CSR format (this is not the case of the reproducer), now I know that this case is not supported currently so I will store all the matrix and use mkl_sparse_spmm.

Thanks.

pd: Sorry about the english

Highlighted
##

MKL 2020 has been released on Dec 12 and this version still doesn't support these matrixes. We only add some note: "Currently, only SPARSE_MATRIX_TYPE_GENERAL is supported.". In the case, if you are interested to extend the supported matrix, please fill the request into Intel Online Service Center.

Gennady_F_Intel

Moderator

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-15-2019
09:45 PM

36 Views

For more complete information about compiler optimizations, see our Optimization Notice.