Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
* add max_dt to prevent too large timestep * adding injection test case for enclosed/open domain * fix signs on divu terms to match paper * Added source terms for rhoH and rhoYm in divu calc needed for injection; running, but something isnt quite right... * adding confined domain pressure correction for mass injection * updates to EnclosedInjection files * add max_dt to prevent too large timestep * adding injection test case for enclosed/open domain * fix signs on divu terms to match paper * Added source terms for rhoH and rhoYm in divu calc needed for injection; running, but something isnt quite right... * adding confined domain pressure correction for mass injection * updates to EnclosedInjection files * Cleaning up and reverting divU sign change * Update EB_ChallengeProblem with new PMF. (#32) * Remove managed memory to enable running with HIP (ROCm 4.5) (#33) * Remove managed arg. from ProbParm * Fix SunMem init. * Add a 3D input file to FlameSheet. * Add timer on SDC k update portion. * Remove mask from inst. RR calculation. EB are handled using box flag and isCovered. * Remove Managed arg from RegTests. * Remove mamged arg. from TripleFlame case. * Add FlameSheet dodecane Lu inputs. (#34) * Fix efield (#35) * Fix Efield BCfill with PMF changes. * Update 1D efield flame source code. * Enable initializing data from pltfile (#36) * Enable initializing data from a pltfile, possibly a PeleLM one as long as the mass fractions are stored in there. * Add Extern/amrdata to the list of AMReX packages. * Tprof regions (#37) * Setup Tprof regions in advance/oneSDC to get clearer profiling data. * Pass MAGMA flag to TPL and add CUDA ptxas optim. off flag. * Switch to MF interpolators. (#38) * Multi-component diffusion (#39) * Add a Multi-Component diffusionOp to class. * add ncomp to diffusionOp constructor, default to 1 * getMCDiffusionOp function. if ncomp != m_ncomp, reset the operator. * Reset mcdiffusionOp during regrid. * Enable DiffusionOp::diffuse_scalar, DiffusionOp::computeDiffFluxes and DiffusionOp::computeDiffLap with multiple components. * Switch to getMCDiffusionOp when dealing with species. * NewStateRedist was removed from AMReX-Hydro * Change level data layout (#40) * Overhaul level data structure, introducing state MF in place of separate MF for velocity, density, species, ... Fix a bug on the averageDownState function that previously did nothing ... * Restore TurnInflow. Change the process slightly where s_ext now contains the turbulence data coming into bcnormal such that the user can decide to overwrite locally the turbulence if needed. * Set level data I_R to zero if we are not doing divu iters or it could get junk. * Adding sumodules to LMeX (#41) * A fair amount of wrongly commited changed in previous PR, but some were good. (#42) * Fix sundials6 (#43) * Rmove call to MemHelper initialize. * Add USE_SUNDIALS as preamble to Make.PeleLMeX. * Fix change in the way umac growth faces are filled. * Update submodules. * Tune down most warnings in LMeX. (#44) * Update perio test (#45) * Update the convected Species/Temperature gaussian test inputs. * Add missing mean flow direction in Convected Gaussian. * Update periodic regtest to unable diffusion convergence test (#46) * Add parameter controling the gaussian width based on a diffusion process. * Add parsing. * Update initalization to properly setup the diffusion problem. * Small change to pprocConvOrder.py to aneble hacking the reference solution. * Add an input file to generate the reference analytical solution. * Fix input.2d_DiffGauS for convergence testing. * Fix turbInflow (#47) * Fix parsing tagging input for box. * Couple of fixes for turbinflow for error introduced while re-arranging the level data. * Fix TurbInflow case to match changes in turbinflow velocity field handling. * Add a input file with refinement on loZ. * Initialize unitTest (#48) * Abort if runMode is not right. * Parse the runMode and evaluate variables list. * Cut the init short if we are in evaluate mode. * Add a container for the evaluate variables. * Add a evaluate capability for a few components of the advance function. Derive can now work without passsing a function ptr, just to keep the same list class. * Add a UnitTest for a dodecane PMF. * Minor tweaks to Evaluate mode. * Fix fill crse patch turb inflow (#49) * Avoid using nan in s_ext when using inflow turb. * Pre-fill the new fine MF with TurbInflow data before filling the patch. * fillcrsepatch with nGrow to avoid using a fillPatch later * Improper definition of turb_center * Add a stiffer scalar advection version, convecting a Gaussian in a periodic box. (#50) * Add M. Reith 1NJOD case. (#51) * Setup 1NJOD case. * Add velocity maximas with verbose >= 2 * If turb is not activated when USE_TURBINFLOW is defined, don't read turbdata. * Something's up with eta, just get rid of turb outside the jet for now. * Remove incomplete minmax print of velocity in computeDt. * Update sub mods (#53) * Updating submodules. * Add missing USE_HIP argument to make TPL * Pass state Array4 in initdata instead of separate Array4 for each entry. (#54) * Update the Case/* pelelm_initdata to take in state Array4 instead of separate entries. * Update call to pelelm_initdata to take in state_arr instead of separate entries. * Update Efield folder. * Update UnitTest folder. * Update RegTest/EB* cases. * Update the remainder of the tests. * Activate fpe in HotBubble regtest 2D. * Add a couple of HoBubble regtests. * Need to keep the aux Array4. * Update submods. * Missing change to state Array4 * Remove deprecated include of pmf.H * Make PeleLM.H trans_parms static. (#55) * Restore the ability to run incompressible flows (#56) * Restore the incompressible run time option. * Update EB_EnclosedVortex with an incompressible run input file. * Add an incompressible decaying HIT case. * Update input to enable running out-of-the-box. * Add HIT to CI suite. * Add a background gp0 option and make gravity of GpuArray. * Add an EB_PipeFlow RegTest. * Add EB_PipeFlow to CI. * Gravity need to be a vector now. * Fix file name in HIT. * Shorten PipeFlow test. * Update TurbInflow (#57) * Remove compile flag from GMake. * Move the turb data into PeleLM.H and update calls to refactored turbinflow. * Update TurbInflow regtest. * Update GMake and input files. * Add TurbInflow tests to CI. * Activate FPE traps while testing TurbInflow * Fix input typo and coarsen turbinflow tests. * Update submodules. * Add support for Bell, Dawson & Shubin (BDS) advection scheme (#58) * Add BDS folder of AMReX-Hydro * For now, add a separate string for scalar advection scheme. * Use m_scalar_advection_scheme for AofS and AofV. Predict velocity still uses Godunov. * Update handling of advection keys to match recent changes in IAMR. * Update 2D input bubble. * Add AA JetInCrossflow case. BDS looks good there. * Update submodule for BDS * Update normal jet case for new TurbInflow. * Add message checking lifted from AmrLevel. (#59) * Add the option to zero visc the diffusivity and thus add the zero_visc (#60) * Add the option to zero visc the diffusivity and thus add the zero_visc kernel to all the pelelm_prob.H * Domain is now used even with EB * Restart plt cpp (#61) * Start extracting the data from plt Header. * Replace amrData by a set of C++ routine using AmrCore FillPatch operations. We can now restart a finer simulation using the pltfile of a coarser one too. * Remove fortran. * PltFileManager moved to PelePhysics. * Update the PltFileManager calls. * Update submodules. * Add flexibililty to EB refinement (#52) * Quick relaxation of EB refinement constrain. * Set up a signed distance function at level 0 to control EB refinement. * Add a function to get the signed distance at any level from the level0 class data. * Parse the EB-refinement strategy keys. * Implement 'Static' EB-refinement strategy. Logic to derefine is still a bit approximative. * Switch the modified FillSignedDistance. Need AMReX PR merge. * Don't change AMReX signDist function and add some code to extend it manually. Might need further testing. * Revert change in capturing the Reactor type. Default is ReactorNull. * Fix FPE while filling level0 signdist * Swirl flow wall interaction (#62) * Add M. Reza case. First version. * Update Mohammadreza case. * Update input file. * Add pmf_data file * Use previous Ubuntu for now, latest has issues with apt-get update * Shorten pipeflow test. * add max_dt to prevent too large timestep * adding injection test case for enclosed/open domain * fix signs on divu terms to match paper * Added source terms for rhoH and rhoYm in divu calc needed for injection; running, but something isnt quite right... * adding confined domain pressure correction for mass injection * updates to EnclosedInjection files * adding injection test case for enclosed/open domain * Added source terms for rhoH and rhoYm in divu calc needed for injection; running, but something isnt quite right... * adding confined domain pressure correction for mass injection * updates to EnclosedInjection files * Cleaning up and reverting divU sign change * removing unused arguments in compute_divu * removine old/new get state call * Removing one more instance of unused variables * Update submodules. * Fix FPE by passing timeStamp to initTemporal function. Add pNew to tempState Update the TurbInflow case with changes in PelePhysics. * addressing PR comments Co-authored-by: Lucas Esclapez <13371051+esclapez@users.noreply.github.com> Co-authored-by: Lucas Esclapez <lucas.esclapez@nrel.gov>
- Loading branch information