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

Showing results for

- Intel Community
- Software
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library
- More Threads with G than S in mkl_dcsrmv?

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

Robert_P_2

Beginner

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

03-30-2014
07:29 PM

75 Views

More Threads with G than S in mkl_dcsrmv?

Link Copied

5 Replies

Chao_Y_Intel

Employee

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

03-30-2014
08:12 PM

75 Views

Hello,

Thanks for your report. By default Intel MKL may choose the threading number dynamically according to some factors, for example, the matrix types, and data size, CPU types. but use can also control the total threading numbers by use the following environment setting:

MKL_DYNAMIC=FALSE

MKL_NUM_THREADS= number of the threadings.

You can set MKL_DYNAMIC=FALSE to check if the symmetric can run more threadings.

Thanks,

Chao

Robert_P_2

Beginner

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

03-31-2014
06:52 AM

75 Views

Chao_Y_Intel

Employee

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

04-13-2014
07:29 PM

75 Views

Hi,

It is possible to be related to the internal implementation. I will check with engineer owner for a few details.

Regards,

Chao

Chao_Y_Intel

Employee

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

04-13-2014
10:16 PM

75 Views

Hi,

We talked with function owner. The current implementation for symmetric matrix-matrix multiplication has some overhead for the small matrix on the large core. So it is suggested to use the non-symmetric interfaces now.

Thanks for checking this.

Regards,

Chao

Robert_P_2

Beginner

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

04-14-2014
10:01 AM

75 Views

This is very helpful, thank you. I will stick to the general version for now.

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

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