examples/fluids: make interface/tests support PETSc main

Merge pull request #801 from CEED/jeremy/no-switches

Refactoring Solids to Remove Switches

Makefile: prereqs for new examples/solids layout

Mooney-Rivlin model (#794)

* refactored MR to work with current version of main. Updated doc with dS for MR

* fixed documentation formatting

* update dS for MR in documentation

* cleaning u

Mooney-Rivlin model (#794)

* refactored MR to work with current version of main. Updated doc with dS for MR

* fixed documentation formatting

* update dS for MR in documentation

* cleaning up MR dS equations and wording

* minor changes in doc

* MR dS draft 1"

* MR coded, jacobian diverging

* added FSInitial-MR1.h

* fixed call option

* NH no longer needs mu_1 option

* processphysics and upated naming scheme

* WIP: Notes on using finite differencing for dS

* added print strain at every increment cl option

* MR cleaned. started CSV file but not quite there

* examples/solids: write csv file

can be read using pandas:

In [2]: pd.read_csv('test.csv')
Out[2]:
increment energy
0 0.000000 0.000000
1 0.066667 0.002402
[...]

* examples/solids: add targets for CSV logging and Altair chart

The MR data is currently hacked because it isn't converging as I'd expect.

* fixed style

* solids/doc: added derivation of dS for MR

* added analytical dS in Jacobian evaluation

* fixed style

* updated doc and energy plot for MR

* MR almost done

* started removing integer division

* got rid of all integer divisions

* missed some 1/3

* remove extra print statement

* added coulped MR FSInitial-MRc qfunction

* MR coupled version doc update

* refactored MR to work with current version of main. Updated doc with dS for MR

* fixed documentation formatting

* update dS for MR in documentation

* cleaning up MR dS equations and wording

* minor changes in doc

* MR dS draft 1"

* MR coded, jacobian diverging

* added FSInitial-MR1.h

* fixed call option

* NH no longer needs mu_1 option

* processphysics and upated naming scheme

* added print strain at every increment cl option

* MR cleaned. started CSV file but not quite there

* examples/solids: write csv file

can be read using pandas:

In [2]: pd.read_csv('test.csv')
Out[2]:
increment energy
0 0.000000 0.000000
1 0.066667 0.002402
[...]

* examples/solids: add targets for CSV logging and Altair chart

The MR data is currently hacked because it isn't converging as I'd expect.

* fixed style

* solids/doc: added derivation of dS for MR

* added analytical dS in Jacobian evaluation

* fixed style

* updated doc and energy plot for MR

* MR almost done

* started removing integer division

* got rid of all integer divisions

* missed some 1/3

* fixed some typo

* rerunning plot files

* Mooney-Rivlin coupled model

* some minor changes

* Mooney-Rivlin coupled model

* some clean-up and update

* clean-up, rebased

* examples/solids: revise docs/derivation and fix Newton

The (fixed) Newton linearization yields an indefinite operator causing
Newton to fail (so output/MR-strain1.csv is broken). This can be handled
by switching to minres, but a better solution would be nice.

* format: added newlines and removed unused/commented code blocks

* doc: removed GP model dropdown.

* examples/solids: Mooney-Rivlin docs/notation

* examples/solids: Mooney-Rivlin notation and experimental configuration

* examples/solids: process -dm_plex_box_faces with PETSc main

* make style

* examples/solids: syntax fix and minor cleanup

* make style

* doc: fixing Holzapfel ref in Mooney-Rivlin

* added TESTARGS for Mooney-Rivlin

* fixed TESTARGS for MR

* tests/tap.sh: make test for solids examples with names

* WIP: adding test option. added -expect_final_strain_energy and RegressionTests_solids function but not sure how to actually pull the reference strain energy from the reference file

* examples/solids: add regression tests using expected strain energy

* examples/solids: error on MMS with model that does not provide manufactured solutions

* examples/solids: fix HIP via CEED_QFUNCTION_HELPER

* update in-line comments

* changed as Jeremy suggested

* examples/solids: make style

Co-authored-by: Rezgar Shakeri <Rezgar.Shakeri@colorado.edu>
Co-authored-by: Jed Brown <jed@jedbrown.org>

show more ...

Merge branch 'release'

Makefile CC_VENDOR: support cross-compilation strings like aarch64-linux-gnu-gcc-10

Merge pull request #770 from CEED/jed/cuda-ci

gitlab-ci: add lv-cuda

make: a bit quieter output

make: consolidate examples/petsc utilities into library to avoid races

Parallel recursive make races itself building the utilities when
creating multiple executables, so consolidate so invocations n

make: consolidate examples/petsc utilities into library to avoid races

Parallel recursive make races itself building the utilities when
creating multiple executables, so consolidate so invocations never build
overlapping targets.

show more ...

Merge pull request #768 from CEED/jeremy/remove-dead

Remove Template Gallery and Backend Code

backends - remove template backends, simplify testing and makefile

Merge pull request #764 from CEED/jeremy/ci-exit

junit - set CEED_ERROR_HANDLER to exit

petsc - clean up some header inclusions

gitlab - add tmpl backends to CPU testing

make - update setting BACKENDS from env, junit was accidentally runing 2x

Merge pull request #746 from CEED/jeremy/python-clean

make - small updates to makefile for python test move

Merge pull request #745 from CEED/jeremy/test-order

gitlab - rearange the script for clearer output

Makefile: fix test for valgrind

make - small updates to makefile for python test move

Makefile: fix typo

Reported-by: Yohann Dudouit <yohann.dudouit@gmail.com>

make configure: add OCCA_DIR

Merge pull request #722 from CEED/jed/no-default-cuda-hip

Disable autodetection of HIP and CUDA

Disable autodetection of HIP and CUDA

Some machines, especially HPC machines with shared filesystems (e.g., at
LLNL) have CUDA and/or HIP compilation stacks despite no such devices.
Autodetection pr

Disable autodetection of HIP and CUDA

Some machines, especially HPC machines with shared filesystems (e.g., at
LLNL) have CUDA and/or HIP compilation stacks despite no such devices.
Autodetection produces confusing errors in such environments, so we
disable them and require explicitly setting CUDA_DIR and/or HIP_DIR to
locate them (picked up automatically from the environment if present).
Document this behavior.

Reported-by: Natalie Beams <nbeams@icl.utk.edu>

show more ...

Merge pull request #716 from CEED/jed/install-backend.h

Jed/install backend.h

Install install backend headers under include/ceed/

This makes it possible to distribute source plugins that provide
additional backends. It's also used in MFEM, perhaps temporarily.

Deprecate ceed

Install install backend headers under include/ceed/

This makes it possible to distribute source plugins that provide
additional backends. It's also used in MFEM, perhaps temporarily.

Deprecate ceed-backend.h, which was not previously installed, but some
users accessed it from an in-place build.

Also install CUDA and HIP headers that allow users to provide CUfunction
and hipFunction_t.

Co-authored-by: Jeremy L. Thompson <jeremy.thompson@colorado.edu>
Requested-by: Andrew T. Barker <barker29@llnl.gov>

show more ...