Using Extrema_ExtAlgo_Tree in ShapeHealing

Damir Bogrov's picture

Hello!

We analyzed shape-healing process in case where this process is a bottleneck. We loaded an invalid shape and performed shape-healing process. On some small files (< 50 faces) this process takes significant time (some secs). The root cause is that projection of 3d curve on surface takes most of the time. As result of our research we found that using a new Extrema_ExtAlgo_Tree algotithm is more efficient than Extrema_ExtAlgo_Grad used by default.
The results in attached pdf-file.
http://www.cadexchanger.com/download/doc/Using_Extrema_ExtAlgo_Tree_in_S...

We would like to ask for your opinion and any information about down-side or side-effects?

Have you considered this algorithm in ShapeHealing and if yes, why have you rejected it?

Best regards,
Damir Bogrov

abv's picture

Please give more details

Hello Damir,

Thank you for sharing your experience!
We have tried using this mode several years ago, but abandoned because it was found slower by that time... perhaps we shall try again.

Can you share some more details: what version of OCCT you are using, what performance gain do you obtain? It would be great if you could share some models allowing to reproduce the effect.

Please note that in OCCT 6.9.0 we have made serious improvements exactly for this functionality, thus I advise you to check your cases on 6.9.0.beta1.

Andrey

Roman Lygin's picture

Slides

Hi Andrey,

All the key details are in the slideset (in .pdf) - OCC version, HW details, performance gains, bottlenecks, etc.
The original models come from customers, we'll check if we can share those.

Thanks for the 6.9.0b1 recommendation.

Roman

P.S. Again, apparently no way to upload the files here on the forum, we had to upload to our website.

Damir Bogrov's picture

testmodels

Hello, Andrey!

Thank you for your qiuck answer!

This is a testmodels:
www.cadexchanger.com/download/doc/testmodels.zip

Damir

abv's picture

Confirm your result

Hello Damir,

I confirm your results, thank you for providing sample data!
I suggest you register an issue in Mantis and provide a fix so that it could be tested on complete test data set.

Andrey

Damir Bogrov's picture

Issue 26166

Tracker has been created [#issue_26166].

www.opencascade.com

Copyright 2011-2017
OPEN CASCADE SAS
Contact us