Turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- Intel Community
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library & Intel® Math Kernel Library
- Symmetric Sparse Inversion

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

vahid_s_

Beginner

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

11-04-2013
01:46 PM

31 Views

Symmetric Sparse Inversion

Hi,

I am trying to find an inverse of a sparse matrix and then multiply it with another sparse matrix: A*inv(B). But I could not find any routine for making inverese of a sparse matrix. The only function that I found is inv() which I don not know is for sparse matrix or not and if it is how can I input the sparse matrix.

Any help is appreciated. Thanks

Link Copied

3 Replies

vahid_s_

Beginner

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

11-04-2013
01:51 PM

31 Views

A and B are symetric and I'm using kernel library in FORTRAN.

Zhang_Z_Intel

Employee

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

11-04-2013
03:07 PM

31 Views

To get inv(B), you can solve the system ** B*inv(B)=I**, where

Also, do you intend *B** *as a sparse upper or lower triangular matrix? If so then it might be easier to use the "mkl_?csrsm" function. This can compute ** inv(B')*A**, then take a transpose of the result to get what you need. See more about "mkl_?csrsm" here: http://software.intel.com/en-us/node/468602#D4C63016-4279-4E47-8ECA-DC04B2161C5E

mecej4

Black Belt

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

11-04-2013
03:21 PM

31 Views

Vahid, what do you intend to do with the inverse, once you have found it?

This is an important question, and several articles exist wherein it is argued that, with a few exceptions, both efficiency and accuracy are sacrificed by forming the inverse explicitly.

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