Showing results for

- Intel Community
- Software
- Software Development SDKs and Libraries
- Intel® oneAPI Math Kernel Library
- 2D Convolution methode

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

KIM__ILTAEK

Beginner

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

04-06-2020
07:23 PM

232 Views

2D Convolution methode

Hi,

Matlab supports three methods when using conv2. As you can see from the reference site below, Full, Same, and Valid methods are supported.

https://johnloomis.org/ece563/notes/filter/conv/convolution.html

Is it possible to add an option like this using vsldConvExec? Or if there is any other way, please advise.

1 Solution

Gennady_F_Intel

Moderator

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

04-08-2020
09:54 PM

232 Views

here is what I see on my part.

Math kernel Lib Runtime : 1.859sec

Naive Algorithm Runtime : 1.157sec

Press any key to continue . . .

Yes, the naive implementation is faster then what mkl suggested. You may submit the request to the intel online service center or as I already suggested you - check the IPP library. The Conv/Corr functionality from IPP is very much optimized.

Link Copied

6 Replies

Gennady_F_Intel

Moderator

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

04-07-2020
08:52 PM

232 Views

KIM__ILTAEK

Beginner

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

04-08-2020
06:40 PM

232 Views

Thank for your reply.

First, we implemented convolution by full method and crop by input size.

However, it is slower than the navie algorithm. I am developing on windows 10 with i7-5960X.

Below is my test code result. Please advise if I have something wrong. Code is attached. Please review.

Math kernel lib runtime : 1.282sec

Naive Algorithm runtime : 0.234sec

Gennady_F_Intel

Moderator

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

04-08-2020
08:14 PM

232 Views

KIM__ILTAEK

Beginner

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

04-08-2020
08:36 PM

232 Views

Please review together.

Gennady_F_Intel

Moderator

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

04-08-2020
09:54 PM

233 Views

here is what I see on my part.

Math kernel Lib Runtime : 1.859sec

Naive Algorithm Runtime : 1.157sec

Press any key to continue . . .

Yes, the naive implementation is faster then what mkl suggested. You may submit the request to the intel online service center or as I already suggested you - check the IPP library. The Conv/Corr functionality from IPP is very much optimized.

KIM__ILTAEK

Beginner

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

04-08-2020
10:04 PM

232 Views

Thank you for your confirm.

I will check it again using IPP.

- 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.