Include openGJK 1.1 #20

MattiaMontanari · 2019-07-29T13:07:41Z

New pull request which addes openGJK 1.1 to Neper devel branch

rquey · 2019-09-25T11:44:59Z

@MattiaMontanari, it does not seem to work as well as the original code.

Sample tesr:
neper -T -n 2 -for tesr

The original code (c3dd37d) converges to f=0:

$ neper -T -n from_morpho -morpho "tesr:file(n2-id1.tesr)" 

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 3.4.1-46
Info   : Built with: gsl nlopt libscotch openmp
Info   : Running on 8 threads.
Info   : <http://neper.sourceforge.net>
Info   : Copyright (C) 2003-2019, and GNU GPL'd, by Romain Quey.
Info   : Comments and bug reports: <[email protected]>.
Info   : Loading initialization file `/home/rquey/.neperrc'...
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] -format tess -tesrformat ascii
Info   : [com line] -n from_morpho -morpho tesr:file(n2-id1.tesr)
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellation...
Info   :     [i] Parsing file `n2-id1.tesr'...
Info   :     [i] Parsed file `n2-id1.tesr'.
Info   :       . Number of cells: 2
Info   :   - Setting seeds... 100%
100%
Info   :   - Generating crystal orientations...
Info   :       . Listing cell voxels... 100.0%
Info   :       . Sampling factor along x = 3.938980.
Info   :       . Sampling factor along y = 3.938980.
Info   :       . Sampling factor along z = 3.938980.
Info   :       . Number of voxels per cell: min = 43, avg = 43, max = 43.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =0.001130725
Info   :     > Iteration    729: fmin=0.000000000 f=0.000000000
Info   :     > Reached `x(r)eps' criterion.
Info   : Writing results...
Info   :     [o] Writing file `nfrom_morpho-id1.tess'...
Info   :     [o] Wrote file `nfrom_morpho-id1.tess'.
Info   : Elapsed time: 0.203 secs.
========================================================================

The pull-request code generates "Initial solution: f =nan" and does not converge to 0:

./neper -T -n from_morpho -morpho "tesr:file(n2-id1.tesr)" 

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 3.4.1-46
Info   : Built with: gsl nlopt libscotch openmp
Info   : Running on 8 threads.
Info   : <http://neper.sourceforge.net>
Info   : Copyright (C) 2003-2019, and GNU GPL'd, by Romain Quey.
Info   : Comments and bug reports: <[email protected]>.
Info   : Loading initialization file `/home/rquey/.neperrc'...
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] -format tess -tesrformat ascii
Info   : [com line] -n from_morpho -morpho tesr:file(n2-id1.tesr)
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellation...
Info   :     [i] Parsing file `n2-id1.tesr'...
Info   :     [i] Parsed file `n2-id1.tesr'.
Info   :       . Number of cells: 2
Info   :   - Setting seeds... 100%
100%
Info   :   - Generating crystal orientations...
Info   :       . Listing cell voxels... 100.0%
Info   :       . Sampling factor along x = 3.938980.
Info   :       . Sampling factor along y = 3.938980.
Info   :       . Sampling factor along z = 3.938980.
Info   :       . Number of voxels per cell: min = 43, avg = 43, max = 43.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =nan
Info   :     > Iteration    102: fmin=0.001267651 f=0.001267651
Info   :     > Reached `eps' criterion.
Info   : Writing results...
Info   :     [o] Writing file `nfrom_morpho-id1.tess'...
Info   :     [o] Wrote file `nfrom_morpho-id1.tess'.
Info   : Elapsed time: 0.024 secs.
========================================================================

MattiaMontanari · 2019-10-09T19:26:22Z

I have pushed a new version as I couldn't find the bug in the previous one... This seems to work now.

rquey · 2019-10-14T08:50:39Z

It works for the previous case (n=2, id=1) but unfortunately not in general. Here is an example with n=4:

Sample tesr:
neper -T -n 4 -for tesr

Version 3.5.0 converges to f=0:

$ neper -T -n from_morpho -morpho "tesr:file(n4-id1.tesr)" -morphooptistop "val=1e-9"

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 3.5.0             
Info   : Built with: gsl, nlopt, openmp
Info   : Running on 8 threads.
Info   : <http://neper.info> <http://neper.info/community>
Info   : Copyright (C) 2003-2019, and GNU GPL'd, by Romain Quey.
Info   : Loading initialization file `/home/rquey/.neperrc'...
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:          
Info   : [ini file] -format tess,geo -tesrformat ascii                    
Info   : [com line] -n from_morpho -morpho tesr:file(n4-id1.tesr)
         -morphooptistop val=1e-9                
Info   : ---------------------------------------------------------------
Info   : Reading input data...               
Info   : Creating domain...
Info   : Creating tessellation...                       
Info   :     [i] Parsing file `n4-id1.tesr'...      
Info   :     [i] Parsed file `n4-id1.tesr'.            
Info   :       . Number of cells: 4                
Info   :   - Setting seeds... 100%
100%                                                                    
Info   :   - Generating crystal orientations...
Info   :       . Listing cell voxels... 100.0%    
Info   :       . Sampling factor along x = 3.126371.
Info   :       . Sampling factor along y = 3.126371.
Info   :       . Sampling factor along z = 3.126371.
Info   :       . Number of voxels per cell: min = 33, avg = 62, max = 100.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =0.002251995
Info   :     > Iteration   5866: fmin=0.000000000 f=0.000000000
Info   :     > Reached `val' criterion.
Info   : Writing results...
Info   :     [o] Writing file `nfrom_morpho-id1.tess'...
Info   :     [o] Wrote file `nfrom_morpho-id1.tess'.
Info   :     [o] Writing file `nfrom_morpho-id1.geo'...
Info   :     [o] Wrote file `nfrom_morpho-id1.geo'.
Info   : Elapsed time: 2.738 secs.
========================================================================

The pull-request code does not converge to 0:

$ neper -T -n from_morpho -morpho "tesr:file(n4-id1.tesr)" -morphooptistop "val=1e-9"

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 3.4.1-46
Info   : Built with: gsl nlopt openmp
Info   : Running on 8 threads.
Info   : <http://neper.sourceforge.net>
Info   : Copyright (C) 2003-2019, and GNU GPL'd, by Romain Quey.
Info   : Comments and bug reports: <[email protected]>.
Info   : Loading initialization file `/home/rquey/.neperrc'...
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] -format tess,geo -tesrformat ascii
Info   : [com line] -n from_morpho -morpho tesr:file(n4-id1.tesr)
         -morphooptistop val=1e-9
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellation...
Info   :     [i] Parsing file `n4-id1.tesr'...
Info   :     [i] Parsed file `n4-id1.tesr'.
Info   :       . Number of cells: 4
Info   :   - Setting seeds... 100%
100%
Info   :   - Generating crystal orientations...
Info   :       . Listing cell voxels... 100.0%
Info   :       . Sampling factor along x = 3.126371.
Info   :       . Sampling factor along y = 3.126371.
Info   :       . Sampling factor along z = 3.126371.
Info   :       . Number of voxels per cell: min = 33, avg = 62, max = 100.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =0.002251996
Info   :     > Iteration  82821: fmin=0.000009213 f=0.000009213
Info   :     > Reached `x(r)eps' criterion.
Info   : Writing results...
Info   :     [o] Writing file `nfrom_morpho-id1.tess'...
Info   :     [o] Wrote file `nfrom_morpho-id1.tess'.
Info   :     [o] Writing file `nfrom_morpho-id1.geo'...
Info   :     [o] Wrote file `nfrom_morpho-id1.geo'.
Info   : Elapsed time: 8.725 secs.
========================================================================

update devel

MattiaMontanari · 2019-10-14T14:13:23Z

Somehow I'm not able to reproduce the results you're showing from this repository (I tried both master v3.5.0 and devel V3.5.1-3). This is what I get from master:

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 3.5.0
Info   : Built with: gsl, nlopt, openmp
Info   : Running on 24 threads.
Info   : <http://neper.info> <http://neper.info/community>
Info   : Copyright (C) 2003-2019, and GNU GPL'd, by Romain Quey.
Info   : No initialization file found (`/home/mmontanari/.neperrc').
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] (none)
Info   : [com line] -n from_morpho -morpho tesr:file(n4-id1.tesr)
         -morphooptistop val=1e-9
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellation...
Info   :     [i] Parsing file `n4-id1.tesr'...
Info   :     [i] Parsed file `n4-id1.tesr'.
Info   :       . Number of cells: 4
Info   :   - Setting seeds... 100%
100%
Info   :   - Generating crystal orientations...
Info   :       . Listing cell voxels... 100.0%
Info   :       . Sampling factor along x = 3.126371.
Info   :       . Sampling factor along y = 3.126371.
Info   :       . Sampling factor along z = 3.126371.
Info   :       . Number of voxels per cell: min = 33, avg = 62, max = 100.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =0.002251995
Info   :     > Iteration   698184: fmin=0.000000855 f=0.000000855
Warning:     > Stopping on Ctrl+C invocation.
Info   : Writing results...
Info   :     [o] Writing file `nfrom_morpho-id1.tess'...
Info   :     [o] Wrote file `nfrom_morpho-id1.tess'.
Info   : Elapsed time: 81.011 secs.
========================================================================

If I run openGJK in the new devel branch I get the initial f and final fmin:

========================    N   e   p   e   r    =======================
Info   : A software package for polycrystal generation and meshing.
Info   : Version 3.5.1-3
Info   : Built with: gsl, nlopt, openmp
Info   : Running on 24 threads.
Info   : <http://neper.info> <http://neper.info/community>
Info   : Copyright (C) 2003-2019, and GNU GPL'd, by Romain Quey.
Info   : No initialization file found (`/home/mmontanari/.neperrc').
Info   : ---------------------------------------------------------------
Info   : MODULE  -T loaded with arguments:
Info   : [ini file] (none)
Info   : [com line] -n from_morpho -morpho tesr:file(n4-id1.tesr)
         -morphooptistop val=1e-9
Info   : ---------------------------------------------------------------
Info   : Reading input data...
Info   : Creating domain...
Info   : Creating tessellation...
Info   :     [i] Parsing file `n4-id1.tesr'...
Info   :     [i] Parsed file `n4-id1.tesr'.
Info   :       . Number of cells: 4
Info   :   - Setting seeds... 100%
100%
Info   :   - Generating crystal orientations...
Info   :       . Listing cell voxels... 100.0%
Info   :       . Sampling factor along x = 3.126371.
Info   :       . Sampling factor along y = 3.126371.
Info   :       . Sampling factor along z = 3.126371.
Info   :       . Number of voxels per cell: min = 33, avg = 62, max = 100.
Info   :   - Running tessellation...
Info   :     > Initial solution: f   =0.002251995
Info   :     > Iteration  24977: fmin=0.000000001 f=0.000000001
Info   :     > Reached `val' criterion.
Info   : Writing results...
Info   :     [o] Writing file `nfrom_morpho-id1.tess'...
Info   :     [o] Wrote file `nfrom_morpho-id1.tess'.
Info   : Elapsed time: 98.156 secs.
========================================================================

To get these, I relaxed some of the tollerances (see d605eee).

Perhaps I need to configure cmake differently?

rquey · 2019-10-15T15:08:28Z

Not sure about your not reproducing the 3.5.0 results.
Your implementation now appears to give much better results and is faster than mine - great!
In the below example, openGJK complains * * OPENGJK REACHED MAXIMUM ITERATION NUMBER ! * * * * * * * *. Could it be fixed (muted?)?

$ neper -T -n 100 -morpho gg -for tesr
$ neper -T -n from_morpho -morpho "tesr:file(n100-id1.tesr)" -morphooptistop "val=1e-9"

MattiaMontanari · 2019-10-17T11:21:41Z

Not sure about your not reproducing the 3.5.0 results.
Your implementation now appears to give much better results and is faster than mine - great!
In the below example, openGJK complains * * OPENGJK REACHED MAXIMUM ITERATION NUMBER ! * * * * * * * *. Could it be fixed (muted?)?
$ neper -T -n 100 -morpho gg -for tesr
$ neper -T -n from_morpho -morpho "tesr:file(n100-id1.tesr)" -morphooptistop "val=1e-9"

All done!

Include openGJK 1.1

9b06202

Push new version

7979263

MattiaMontanari and others added 2 commits October 14, 2019 11:32

Merge pull request #2 from rquey/devel

c210fec

update devel

Relax tolerances

d605eee

Clean up: fix headers, remove old functions and silence warnings

bc6dbbe

rquey added the enhancement label Oct 17, 2019

rquey merged commit bc6dbbe into neperfepx:devel Oct 17, 2019

VarmaNarendra mentioned this pull request Jun 25, 2021

Aborted (core dumped) During 2D Meshing #254

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Include openGJK 1.1 #20

Include openGJK 1.1 #20

Uh oh!

MattiaMontanari commented Jul 29, 2019

Uh oh!

rquey commented Sep 25, 2019

Uh oh!

MattiaMontanari commented Oct 9, 2019

Uh oh!

rquey commented Oct 14, 2019

Uh oh!

MattiaMontanari commented Oct 14, 2019

Uh oh!

rquey commented Oct 15, 2019

Uh oh!

MattiaMontanari commented Oct 17, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Include openGJK 1.1 #20

Include openGJK 1.1 #20

Uh oh!

Conversation

MattiaMontanari commented Jul 29, 2019

Uh oh!

rquey commented Sep 25, 2019

Uh oh!

MattiaMontanari commented Oct 9, 2019

Uh oh!

rquey commented Oct 14, 2019

Uh oh!

MattiaMontanari commented Oct 14, 2019

Uh oh!

rquey commented Oct 15, 2019

Uh oh!

MattiaMontanari commented Oct 17, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants