* Support array containers in core grudge DG ops

* Convert wave example to use array containers

* Update projection routine to support array containers

* Convert mpi communication

* Use DOFArray-based traversal routines

* Containerize distributed (eager) MPI communication

* Use array container in wave-op-mpi example

* Clean up wave-op-mpi demo

* Use non-recursive array traversal

* Use map_array_container in trace pair exchange

* Remove unused imports

* Remove sketchy assert; shape differs when evaluating fluxes vs storing state

* Refactor containerization routines

* Fix suppressed logging error

* Fix forward flattening map

* Fix type check in projection routine

* dofdesc-itize descriptors first in projection

* Make flattening methods private and provide short docs

* Clean up flatten/unflatten functions and document

* Carefully track ordering of data in flattened array

* Use container flatten/unflatten functions from arraycontext

* Containerize elementwise reductions and add unit tests

* Containerize nodal reductions and add unit tests

* Update documentation to reflect container support

* Catch scalars in op.norm

* Remove flat_norm and use op.norm instead

* Separate initialization and completion of send/recv

* Fix documentation in operator module

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Rename boundary communicator data/attributes

* Comment on nonblocking MPI communication process in communicator class

* Clean up comments related to rank-communication for trace pairs

* Clean up/clarify documentation related to array containers

* Update reduction unit tests

* Documentation updates

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Fix comment formatting

* Containerize grad/div functions

* Add checks for empty arrays in grad/div operators

* Fix gradient op for vector-valued components

* Correct div helper for handling empty and nested obj arrays

* Clear up documentation and add clarifying notes

* More shared code in array container support (#191)

* Refactor grad/weak grad to use the same helper, simplify div empty handling

* Factor out some redundant code in elementwise reductions

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* reimplement norm() without recursion

* disallow norm of a scalar; fix tests

* Fix diff property, ugh.

* Add missing docstring support

* Test the values, not the truthy property of these arrays.

* Get op (woops).

* Bump tolerance for test_mass_mat_trig in test_grudge_sym_old

* Set tolerances in test_mass_trig (for old sym code) equal to ones in new-style tests

Andreas Klöckner authored 3 years ago and Andreas Klöckner committed 3 years ago

https://github.com/inducer/meshmode/pull/225

* Update function: dt_non_geometric_factor -> dt_non_geometric_factors

* Add timestep estimation helper function

* Add clarifying remark about timestep estimation

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Make wavespeed argument option

* Drop dt estimator; provide characteristic lengthscales

* Expand and update documentation

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Allow HyperbolicOperator.max_characteristic_velocity to return local data

* Add warning to dt_geometric_factors about non-volume discretizations

* Add fudge factor to surface advection time step

* Time step estimation: clean up freeze/thaw semantics

* Wave models: clean up freeze/thaw

* Variable coefficient wave model: do not nodal-reduce characteristic velocity

* Add missing freeze(...) for dim=1 case

* Add missing thaw in dt utils unit test

* Fix variable coefficient wave example

* Memoize characteristic lengths rather than geometric factors

* Use actx from wave speed DOFArray in variable coefficient operator

* Pass array context to estimate_rk4_timestep

* Ensure application actx is passed through to timestep estimation

* Update Operator method interfaces for timestepping

* Always return thawed characteristic lengthscales

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Add jump property to TracePair.

* Change jump --> diff

* Add simple smoke test for trace pair

* placate flake8

* Undo spurious chage to primitives.py

* Tweak TracePair docstrings

Co-authored-by: Andreas Kloeckner <inform@tiker.net>

* Make nodal reductions mpi-aware

* Add clarifying remark for nodal reductions

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Use timestep estimation in mpi wave test

Co-authored-by: Andreas Klöckner <inform@tiker.net>

* Simplify WADG mass inverse

* Test mass inverse on warped geometry

* Always WADG all the time also for symbolic

* Fold test_nonaffine_mass_operator_inverse into test_mass_operator_inverse

Co-authored-by: Thomas Gibson <gibsonthomas1120@hotmail.com>

- Rename, document the supporting functions for improved clarity
- Drop dim argument from (former) surface_normal

The function registry bits are only relevant for symbolic operators,
and the broadcasting test duplicates what's already being done in
arraycontext.

This ensures that the symbolic stuff remains tested until we remove it.

* vectorize local_grad and weak_local_grad

* added stack argument to local_grad/weak_local_grad

* add test for gradient using op

* propagate stack argument

* tweak comment to avoid flake8 warning

* stack -> (not) nested

* add visualization to gradient test

* disable strong form gradient test for now

* minor cosmetic changes

* tweak local_div/weak_local_div vectorization to work with both nested and non-nested arrays

* suppress pylint false positives

* remove inverse_mass around grad/div in strong form tests

* add check for valid 'form' argument

* remove unneeded flux terms from strong form tests

* Tweak phrasing: "no jumps" instead of "smooth"

Co-authored-by: Andreas Klöckner <inform@tiker.net>