Skip to content

Enable support for large kernels in morphological ops #2852

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
umar456 merged 5 commits into from
Apr 22, 2020
Merged

Enable support for large kernels in morphological ops #2852

umar456 merged 5 commits into from
Apr 22, 2020

Conversation

@9prady9
Copy link
Member

@9prady9 9prady9 commented Apr 21, 2020

@9prady9 9prady9 force-pushed the issue1941_morphological_ops branch 2 times, most recently from 4cfb4f5 to 6c7a51d Compare April 21, 2020 12:30
@9prady9
Copy link
Member Author

9prady9 commented Apr 21, 2020 •
edited
Loading

I am looking into the regression failures on windows/ubuntu(MKL) CPU jobs.

@9prady9
Copy link
Member Author

9prady9 commented Apr 21, 2020

Was able to reproduce, seems to be happening with MKL based fft only. Looking into it.

@9prady9 9prady9 force-pushed the issue1941_morphological_ops branch 2 times, most recently from d515c05 to 479bab0 Compare April 22, 2020 09:36
@9prady9 9prady9 requested a review from umar456 April 22, 2020 13:59
@9prady9 9prady9 mentioned this pull request Apr 22, 2020
Merged
9prady9 added 5 commits April 22, 2020 22:45
@9prady9
Use persistent boost env var in windows github ci job
9a942d2
@9prady9
Remove faulty cpu::BinOp struct implementation
d3f39bc
@9prady9
Use backend agnostic flip at af_flip entry
3b06aed 
This removes redundant flip implementation at src/api/c/ level again.
@9prady9
Return input if pad output dims match input
54c8c35
@9prady9
Remove kernel size limit for b8 inputs of erode/dilate
c1d7088 
b8(binary images) don't have any size limitations for
structuring-element/kernel starting with this change. For such
larger kernels, convolution(fft) based implementation is used.
@9prady9 9prady9 force-pushed the issue1941_morphological_ops branch from 479bab0 to c1d7088 Compare April 22, 2020 19:13
@umar456 umar456 merged commit 647cf39 into arrayfire:master Apr 22, 2020
@9prady9 9prady9 deleted the issue1941_morphological_ops branch April 24, 2020 05:33
@umar456 umar456 added this to the 3.7.2 milestone Jun 27, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@umar456 umar456 umar456 approved these changes

Assignees

No one assigned

Labels

feature

Projects

None yet

Milestone

3.7.2

Development

Successfully merging this pull request may close these issues.

Add support bigger CUDA kernel size for Morphological Operations (erode/dilate)

2 participants

@9prady9 @umar456