Ken/idl (#1138)

* inflow sponge pushing p towards P0 compiles and runs but unknown yet whether it is effective or not

* looks like my first attempt at this is missing something as there is neglig

Ken/idl (#1138)

* inflow sponge pushing p towards P0 compiles and runs but unknown yet whether it is effective or not

* looks like my first attempt at this is missing something as there is negligible effect from Jacobian addition

* examples/fluids: IDL fixes and demo

* Fix IDL sign error and rename variables/clang-format
* Let newtonianwave.yaml use -reference_pressure and temperature
* Propagate P0 = reference.pressure

* Small tweak to add IDL to the 3 ramp parameters.

* fluids: Extract fluids binary vector load function

- Needs to be used in multiple places (for ICs and for testing data)

* fluids: Correct typo in previous commits

* fluids: Add IDL and newtonian wave test

* fluids: Rename Newtonian Wave --> Gaussian Wave

* fluids: Move RampCoefficient into utils

- Also change name to LinearRampCoefficient, as we may have different
kinds of ramps in the future

* fluids: Change IDL commandline options

* docs(fluids): Document internal damping layer

Co-authored-by: Jed Brown <jed@jedbrown.org>

* fluids: Add ScaleN QF utility

* fluids: Extract InternalDampingLayer QFunction helper

---------

Co-authored-by: Jed Brown <jed@jedbrown.org>
Co-authored-by: KennethEJansen <jansenke@colorado.edu>
Co-authored-by: James Wright <james@jameswright.xyz>

show more ...

Merge pull request #1133 from CEED/jed/fluids-riemann-outflow

Switch outflow from pressure to Riemann pressure + temperature

examples/fluids: Riemann outflow using pressure and temperature

* This allows stable sustained and transient recirculation.

* short-wake domain recirculation behind cylinder has sustained
rec

examples/fluids: Riemann outflow using pressure and temperature

* This allows stable sustained and transient recirculation.

* short-wake domain recirculation behind cylinder has sustained
recirculation. If the exterior temperature is different, the
recirculation will bring in the wrong temperature, but at least
it's stable.
* nwave with slow background causes transient flow reversal when
strong acoustic hits boundary.
* Clean outflow and lid for Blasius

* I tried the 2 P_ext - P_int strategy from Mengaldo (constant density)
and with specified exterior temperature, but they disrupted the interior
flow and were not significantly less reflective.

* The methods I tried for extrapolating interior temperature were not
stable. Simple extrapolation of velocity worked best, and did not show
signs of instability, perhaps because the neighboring flow and implicit
solve ties it together.

* This commit adds support for softmax control of recirculation, but the
parameters were finnicky, solver nonlinearity was a concern, and I'm not
convinced it's ever better than just allowing recirculation.

show more ...

Vortex shedding / flow past cylinder (#1082)

* add vortex shedding problem

* add file for the Qfunctions

* add yaml file

* update problem setup

* update IC qfunc

* update problem set

Vortex shedding / flow past cylinder (#1082)

* add vortex shedding problem

* add file for the Qfunctions

* add yaml file

* update problem setup

* update IC qfunc

* update problem setup

* update initial condition

* update yaml file

* add mesh file

* add vortexshedding to navierstokes.h

* register vortexshedding to cloptions.c

* update configuration/yaml file

* fix Min function

* examples/fluids: fix boundary condition labels for cylinder

$PhysicalNames
6
2 1 "inlet"
2 2 "outlet"
2 3 "top"
2 4 "bottom"
2 5 "cylinderwalls"
2 6 "frontandback"
$EndPhysicalNames

* update QFunctions in navierstokes.c

* update data and BC structs in navierstokes.h

* update vortexshedding.c

* update states

* update IC in vortexshedding.h

* add freestream BC to cloptions

* updates in setuplibceed.c

* include freestream BC in example

* cleanup duplicate variable names

* rename variables as freestream

* clean-up some warning messages

* clean up example

* add cylinder.geo

* label physical volume

* examples/fluids: fix cylinder mesh and add vortexshedding defaults

* update gravity vector

* examples/fluids: match domain to Shakib, get vortex street

* update to quadratic mesh

* clean up yaml file

* delete vortexshedding.c and vortexshedding.h

* remove vortexshedding from register

* update yaml file

* examples/fluids: Newtonian "still" initial condition

Reinstate outflow boundaries that were previously removed so that
existing tests can pass.

* examples/fluids: freestream inherits reference state

Fix freestream context reference counting bug.

* examples/fluids: avoid designated initializer in kernel code (breaks CUDA)

* add Shakib91 to references

* remove cylinder.geo and cylinder.msh

* wip:add documentation for vortex shedding

* update path to cylinder.msh

* add subdirectory for meshes

* parameterize cylinder.geo

* some fixes in cylinder.geo

* Fix cylinder.geo file

* examples/fluids: fix up mesh parametrization

Fix cylinder size (the function is Sqrt, not sqrt) and parametrize
resolution and aspect.

* examples/fluids: fix build after broken merge from main

* update Shakib1991 in references

* wip:docs for vortex shedding

* remove TODO comment

* reformulate docs

* fix docs

* delete cylinder.msh file

* work on documentation

* add vortex runtime options

* examples/fluids: update vortex shedding docs and mesh

Co-authored-by: Jed Brown <jed@jedbrown.org>

show more ...

Merge pull request #1104 from CEED/jeremy/formatting

Minor Formatting Fixes

minor - assorted formatting fixes

Merge pull request #1102 from CEED/jrwrigh/re-pretty_fluids

style: Prettyify fluids after clang-format switch

style: Prettyify fluids after clang-format switch

Mainly rearrange the input/output variable declarations for QFunctions

Merge pull request #1095 from CEED/jed/remove-stale-indent-comments

remove stale indent comments

style: remove obsolete (with clang-format) INDENT comments

Switch to clang-format (#1051)

* style - switch to clang-format

* ci - use newer libxsmm

* action - update format action

* format - consistent use of {} for multi-line if/for

* make - re

Switch to clang-format (#1051)

* style - switch to clang-format

* ci - use newer libxsmm

* action - update format action

* format - consistent use of {} for multi-line if/for

* make - remove stray newline

* make - simpler 'make format' target

* ci - use newer libxsmm

* doc - minor release note claification

* minor - minor fix

* minor - minor fix

* minor - minor fix

* minor - minor fix

* make format

* format - less aggressive alignment rules

* tidy - check for argument name mismatches

* fix newline

* format - mirror Ratel update to .clang-format

* fix merge error

* fix merge conflict

* fix merge error

* drop style in .phony list

* Update .clang-format

Co-authored-by: Jed Brown <jed@jedbrown.org>

* apply updated format

Co-authored-by: Jed Brown <jed@jedbrown.org>

show more ...

Merge pull request #1080 from CEED/ade/riemann-bcs

HLL Flux for Riemann BCs

examples/fluids: remove stray ceed/types.h (breaking GPU)

fluids: Integrate freestream bc

fluids: Create freestream bc qfunction

style(fluids): Visual cleanup of the fluids code

Also satisfy astyle for RiemannFlux_HLL commit

fluids: Add RiemannFlux_HLL helper function

Co-authored-by: Jed Brown <jed@jedbrown.org>

fluids: Unify Newtonian Primitive functions #1035

Merge separate IFunction and IJacobian routines into a single helper function and
separate wrappers for each state variable option.

fluids: Use newtonian with StateFromQi*_t func pointers

This to "unify" the primitive and conservative formulations of the
newtonian solver. By passing function pointers to a helper function, the
co

fluids: Use newtonian with StateFromQi*_t func pointers

This to "unify" the primitive and conservative formulations of the
newtonian solver. By passing function pointers to a helper function, the
compiler can optimize for a given state function(s).

show more ...

fluids: Use StateFromQi* for newt boundary QFs

fluids: Add FluxTotal_Boundary, Refactor newt BI QFs

- Refactoring the newtonian boundary integral QFunctions

fluids: Rename primitive booleans, misc formatting

QF headers for typedefs and macros (#1036)

* jit - qf headers for typedefs and macros

* jit - smaller list of permitted files

* ceed - only include ceed.h in QF source

Fluids - Minor Refactor (#995)

* Fluids - remove *.bin* since it exists in libCEED/.gitignore

* cleanup

* WIP: Fluids - use FluxInviscid_fwd() for computing the flux jacobean

strong_conv is

Fluids - Minor Refactor (#995)

* Fluids - remove *.bin* since it exists in libCEED/.gitignore

* cleanup

* WIP: Fluids - use FluxInviscid_fwd() for computing the flux jacobean

strong_conv is not reproduced and stab is not computed correctly

* WIP: working on variational notation for stabilization

* Fluids - fix missing parts after rebasing onto main

* Fluids - fix SU in the IFunction

* Fluids - refactor stabilization

* Fluids - move stab helper functions to stabilization.h

* Fluids - minor style

* Fluids - minor refactor

* Fluids - add regression test for primitive variables

* Fluids - use UnpackState_*()

* Fluids - more refactoring of stabilization

* Fluids - DC: use utility functions

* Fluids - refactor total flux

* Fluids - minor cleanup

* Fluids - set x[0] = 0 for channel to exclude gravitational potential

* Fluids - updated stabilization.h

* Fluids - zero strong_conv before pass it to FluxInviscidStrong() to be safe

* Fluids - move switch-case for STAB to a helper function

* Fluids - rename Y -> R in Stabilization*()

* Fluids - fix stab sign in RHS

Co-authored-by: Jed Brown <jed@jedbrown.org>

show more ...

Merge pull request #1029 from CEED/jrwrigh/stg-primitive

fluids: Make Strong STG compatible with primitive variables.