Adding Quadratic elements to Silo abstract knowledge pool

[cessenat]

This adds the quadratic elements ids for Silo.
Next step shall be to make them properly readable in VisIt Silo plugin.
Further step shall be to take a decision about extfaces.c : degenerate quadratic elements down to linear on boundaries so that very few work shall need to be done, or preserve the quadratic feature. I would opt for the first option, what do you think @markcmiller86 ?
Another step shall be for mixed materials ; I think mixed materials is a typical Finite Volume issue, so high order Finite Elements are orthogonal to that question ; so, as for external faces, one could degenerate quadratic to linear for mixed materials.

[markcmiller86]

Be sure to run make silo.inc; make silo_f77.inc; and silo_f9x.inc in src/silo to regen those fortran include files and include those updated files in the PR.

[markcmiller86]

We'll also have to adjust the README.md and some of the docs/*.md files too but we do this in a sep. PR on the ghpages branch.

[markcmiller86]

@rrsettgast if you have some time, can you take a quick look at this and the associated VisIt work and share anything that might come to mind as relevant?

[markcmiller86]

Another step shall be for mixed materials ; I think mixed materials is a typical Finite Volume issue, so high order Finite Elements are orthogonal to that question ; so, as for external faces, one could degenerate quadratic to linear for mixed materials.

Ok, I am looking into the mixed material question. It is def. a volume sort of thing so you may be right that the two really don't combine in any coherent way. OTOH, if every high order element has a linear approx. using just the first set of (linear) dofs, then maybe that is a way that material volume fractions can be specified.

[markcmiller86]

@cessenat I think this looks good. Have you run Silo's test suite against it just to make sure. Typically, just a make check from the top level should do it.

[cessenat]

@cessenat I think this looks good. Have you run Silo's test suite against it just to make sure. Typically, just a make check from the top level should do it.

Tests fail, with or without my modifications. There is something basically wrong with my Ubuntu 20.04.
Downloading https://github.com/LLNL/Silo/archive/refs/tags/4.11.1.tar.gz

cessenat@jliana:/dev/shm/SiloTest/Silo-4.11.1$ ./configure --prefix=/opt/silo-4.11.1 -with-hdf5=/usr/include/hdf5/serial,/usr/lib/x86_64-linux-gnu/hdf5/serial --enable-fortran=yes --with-readline  FC=gfortran CXX=c++ CC=gcc
...
cessenat@jliana:/dev/shm/SiloTest/Silo-4.11.1/tests$ make check

NOTICE: Silo test cases are not compiled by default.  To compile and test
NOTICE: them run "make check" from the silo "tests" directory. Otherwise
NOTICE: you can compile them individually with "make TESTNAME".  The test
NOTICE: names are: array, dir, extface, multi_test, partial_io, point, quad,
NOTICE: simple, ucd, ucdsamp3, testall, obj, onehex, oneprism, onepyramid,
NOTICE: onetet, subhex, TestReadMask, twohex, multispec, misc, sami,
NOTICE: newsami, specmix, spec, cpz1plt, group_test, listtypes, alltypes,
NOTICE: wave, multi_file, polyzl, csg, rocket, mmadjacency, largefile,
NOTICE: dbversion, namescheme, efcentering, mk_nasf_pdb, ioperf, arbpoly2d,
NOTICE: readstuff, mat3d_3across, merge_block, test_mat_compression,
NOTICE: bcastopen, memfile_simple, empty, majorder, realloc_obj_and_opts,
NOTICE: testpdb, pdbtst, compression, grab, mk_nasf_h5, testhdf5, arrayf77,
NOTICE: arrayf90, curvef77, matf77, pointf77, quadf77, ucdf77, testallf77,
NOTICE: csgmesh, qmeshmat2df77, but "testall" is a good one to start with.

make  rocket_silo.la ioperf_silo.la ioperf_pdb.la ioperf_sec2.la ioperf_stdio.la ioperf_hdf5.la arbpoly arbpoly2d array extface multi_test partial_io point group_test listtypes simple ucd ucdsamp3 testall alltypes obj onehex oneprism onepyramid onetet TestReadMask subhex twohex multispec specmix misc sami newsami spec cpz1plt merge_block mat3d_3across ucd1d wave multi_file polyzl csg rocket mmadjacency largefile dir quad dbversion namescheme efcentering mk_nasf_pdb ioperf readstuff testfs empty majorder realloc_obj_and_opts test_mat_compression bcastopen memfile_simple testpdb pdbtst  compression grab mk_nasf_h5 testhdf5 arrayf77 arrayf90 curvef77 matf77 pointf77 quadf77 ucdf77 testallf77 csgmesh qmeshmat2df77 \
  checksums testhzip testxvers test_error.py test_read.py test_write.py onehex.py testonehex testsilock testdtypes xvers_data.tar.xz pion0244.silo zeros.dat
make[1]: Entering directory '/dev/shm/SiloTest/Silo-4.11.1/tests'
make[1]: 'rocket_silo.la' is up to date.
make[1]: 'ioperf_silo.la' is up to date.
make[1]: 'ioperf_pdb.la' is up to date.
make[1]: 'ioperf_sec2.la' is up to date.
make[1]: 'ioperf_stdio.la' is up to date.
make[1]: 'ioperf_hdf5.la' is up to date.
make[1]: 'arbpoly' is up to date.
make[1]: 'arbpoly2d' is up to date.
make[1]: 'array' is up to date.
make[1]: 'extface' is up to date.
make[1]: 'multi_test' is up to date.
make[1]: 'partial_io' is up to date.
make[1]: 'point' is up to date.
make[1]: 'group_test' is up to date.
make[1]: 'listtypes' is up to date.
make[1]: 'simple' is up to date.
make[1]: 'ucd' is up to date.
make[1]: 'ucdsamp3' is up to date.
make[1]: 'testall' is up to date.
make[1]: 'alltypes' is up to date.
make[1]: 'obj' is up to date.
make[1]: 'onehex' is up to date.
make[1]: 'oneprism' is up to date.
make[1]: 'onepyramid' is up to date.
make[1]: 'onetet' is up to date.
make[1]: 'TestReadMask' is up to date.
make[1]: 'subhex' is up to date.
make[1]: 'twohex' is up to date.
make[1]: 'multispec' is up to date.
make[1]: 'specmix' is up to date.
make[1]: 'misc' is up to date.
make[1]: 'sami' is up to date.
c++ -DHAVE_CONFIG_H -I. -I..  -I./../src/silo -I./../src/silo -I/opt/silo-4.11.1/include -I/usr/include/hdf5/serial   -fPIC -MT newsami.o -MD -MP -MF .deps/newsami.Tpo -c -o newsami.o newsami.cxx
In file included from /usr/include/c++/9/tuple:39,
                 from /usr/include/c++/9/bits/stl_map.h:63,
                 from /usr/include/c++/9/map:61,
                 from newsami.cxx:60:
./array:1:1: error: stray '\177' in program
    1 | �ELF         � >    ��      @       ��b         @ 8   @ ( ' �   �   @       @       @       �      �            �   �   ��      ��      ��      �       �                 �                            �       �       �         �    �       �       �      �h�     �h�      �         �     �       �       �     . �     . �      �         �   ���     �(�     �(�     �7      P�      �         �   �&�     �6�     �6�     @      @            �   �   8�      8�      8�                            �   �   X�      X�      X�      D       D       �       S�td�   8�      8�      8�                            P�td�    ��      ��      ��     4M      4M      �       Q�td�                                           �       R�td�   ���     �(�     �(�     0�      0�             /lib64/ld-linux-x86-64.so.2     �   �   �   GNU   ��   �       �   �   �   GNU �?݄ ���V�O.})Vi��o:�   �      GNU     �              C   �         ��@���B�@ �
      | ^
.. bla bla bla...
./array:56984:127: error: ‘H’ does not name a type
In file included from /usr/include/c++/9/tuple:39,
                 from /usr/include/c++/9/bits/stl_map.h:63,
                 from /usr/include/c++/9/map:61,
                 from newsami.cxx:60:
./array:56984:2176: error: expected declaration before ‘}’ token
./array:56984:2215: error: expected class-name before numeric constant

[markcmiller86]

Tests fail, with or without my modifications. There is something basically wrong with my Ubuntu 20.04.

Ok, I think I've seen this kind of failure before. Don't worry about it for now.