Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.
Announcements
The Intel sign-in experience is changing in February to support enhanced security controls. If you sign in, click here for more information.
6680 Discussions

How to search the eigenvalue interval of a matrix by MKL library?

HUANGK
Novice
291 Views

Hi!

Hello, I want to use FEAST to solve a generalized eigenvalue problem, but I have a problem. In FEAST, the eigenvalue interval and the number of eigenvalues of the sparse matrix need to be searched. I checked the FEAST manual and did not find how to search the eigenvalue interval of the matrix. Am I missing something? Is there any built-in function in MKL to calculate the eigenvalue interval of a large sparse matrix?
Your expertise is invaluable to me, thank you for your help!

0 Kudos
1 Solution
VidyalathaB_Intel
Moderator
255 Views

Hi,

 

Thanks for reaching out to us.

 

In order to use the Extended eigensolver interface in MKL, we need to pass the search interval i.e. the lower and upper bounds of the interval to be searched for eigenvalues.

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/exten...

But the largest and smallest eigenvalues can be calculated.

In the MKL documentation, under the section > "Extended eigensolver interface to find the largest and smallest eigenvalues", there are two routines that help in calculating the largest and smallest eigenvalues.

 

mkl_sparse_?_ev > for standard eigenvalue problem

Reference:

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/exten...

mkl_sparse_?_gv > for a generalized eigenvalue problem

Reference:

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/exten...

Please go through the above links for full details and you can also find examples of the respective routines in the mkl folder which helps in better understanding. 

Kindly let us know if this is not what you are looking for by elaborating a bit more on the issue (maybe through an example).

 

Regards,

Vidya.

 

View solution in original post

4 Replies
VidyalathaB_Intel
Moderator
256 Views

Hi,

 

Thanks for reaching out to us.

 

In order to use the Extended eigensolver interface in MKL, we need to pass the search interval i.e. the lower and upper bounds of the interval to be searched for eigenvalues.

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/exten...

But the largest and smallest eigenvalues can be calculated.

In the MKL documentation, under the section > "Extended eigensolver interface to find the largest and smallest eigenvalues", there are two routines that help in calculating the largest and smallest eigenvalues.

 

mkl_sparse_?_ev > for standard eigenvalue problem

Reference:

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/exten...

mkl_sparse_?_gv > for a generalized eigenvalue problem

Reference:

https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/exten...

Please go through the above links for full details and you can also find examples of the respective routines in the mkl folder which helps in better understanding. 

Kindly let us know if this is not what you are looking for by elaborating a bit more on the issue (maybe through an example).

 

Regards,

Vidya.

 

VidyalathaB_Intel
Moderator
220 Views

Hi Huang,

 

Reminder:

 

As we haven't heard back from you, could you please let us know if there is any update regarding your issue? Please do let us know if the provided information helps.

 

Regards,

Vidya.

 

HUANGK
Novice
208 Views
VidyalathaB_Intel
Moderator
198 Views

Hi Huang,


Thanks for accepting our solution.

As the issue is resolved we are closing this thread. Please post a new question if you need any additional assistance from Intel as this thread will no longer be monitored.


Have a Great Day!


Regards,

Vidya.


Reply