Skip to content

imgproc: use universal intrinsic in accumulate series#7110

Merged
opencv-pushbot merged 1 commit intoopencv:masterfrom
tomoaki0705:featureUniversalIntrinsicAccumulate
Aug 15, 2016
Merged

imgproc: use universal intrinsic in accumulate series#7110
opencv-pushbot merged 1 commit intoopencv:masterfrom
tomoaki0705:featureUniversalIntrinsicAccumulate

Conversation

@tomoaki0705
Copy link
Copy Markdown
Contributor

@tomoaki0705 tomoaki0705 commented Aug 14, 2016

This pullrequest changes

  • use universal intrinsic instead of raw SSE/NEON implementation
  • performance comparison is shown below
  • for windows performance measurement, IPP has been switched off to use the universal intrinsic implementation
  • please note that this modification is not for performance improvement, so some "after" numbers are slightly worse than "before" number, but basically similar value
Platform Windows Windows ARM(32bit) ARM(32bit) ARM(64bit) ARM(64bit)
implementation before after before after before after
Video_Acc.accuracy 291 ms 256 ms 961 ms 870 ms 626 ms 623 ms
Video_AccSquared.accuracy 220 ms 214 ms 1016 ms 926 ms 642 ms 643 ms
Video_AccProduct.accuracy 221 ms 222 ms 1027 ms 931 ms 684 ms 685 ms
Video_RunningAvg.accuracy 223 ms 222 ms 997 ms 900 ms 645 ms 648 ms

Platform information

  • Windows: MacBookPro Mid2012 (Windows 7 x64 + VS2012 Update 4 + Corei7 4Core 2.6GHz)
  • ARM 32bit: Jetson TK1(Ubuntu 14.04 + gcc 4.8.4 + ARM Cortex A15 4Core 2.3GHz)
  • ARM 64bit: ODROID-C2 (Ubuntu 16.04 + gcc 5.4.0 + ARM Cortex A53 4Core 2.0GHz)

@tomoaki0705 tomoaki0705 force-pushed the featureUniversalIntrinsicAccumulate branch from 44c6ce2 to 360d31e Compare August 15, 2016 02:26
@tomoaki0705
Copy link
Copy Markdown
Contributor Author

tomoaki0705 commented Aug 15, 2016

I realized that the number from ARM(32bit) and ARM(64bit) are opposite.
Opposite for both before and after, so not a big deal.

use universal intrinsic in accumulate series
4f7679c 
      * accumulate, accumulateSquare, accumulateProduct and accumulateWeighted
@tomoaki0705 tomoaki0705 force-pushed the featureUniversalIntrinsicAccumulate branch from 360d31e to 4f7679c Compare August 15, 2016 04:01
@alalek
Copy link
Copy Markdown
Member

alalek commented Aug 15, 2016

Well done! 👍

@opencv-pushbot opencv-pushbot merged commit 4f7679c into opencv:master Aug 15, 2016
opencv-pushbot pushed a commit that referenced this pull request Aug 15, 2016
@alalek
Merge pull request #7110 from tomoaki0705:featureUniversalIntrinsicAc…
e884bba 
…cumulate
@tomoaki0705 tomoaki0705 deleted the featureUniversalIntrinsicAccumulate branch August 15, 2016 21:35
@tomoaki0705 tomoaki0705 mentioned this pull request Aug 26, 2016
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

@alalek alalek

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

3 participants

@tomoaki0705 @alalek @opencv-pushbot