Repository navigation
Dependencies
- AMReX: 26.11
- pyAMReX: 26.11
- PICSAR-QED: 26.11
- pybind11: 3.1.0
- picmistandard: 0.35.0
- openPMD-api: 0.17.0
- Python: 3.11 - 3.14
What's Changed
- CMake: keep derived AMReX options in sync on re-configure by @ax3l in #7204
- RZ: fix azimuthal-mode parity when folding
rhoatr=0by @RemiLehe in #7331 - Doc: Darwin Solver by @ax3l and @roelof-groenewald in #7306
- refactor implicit initial guess for readability by @JustinRayAngus in #7259
- Use anisotropic chi for Darwin preconditioner by @roelof-groenewald in #7239
- AGENTS.md: use DevUtils script to reset checksums from CI by @RemiLehe in #7332
- add mass matrices for 1D RSPHERE. by @JustinRayAngus in #7312
- Dependencies: weekly update by @github-actions[bot] in #7342
- MultiFabRegister.H : fix bug potentially leading to infinite recursive call and segfault by @lucafedeli88 in #6699
- Use smoother bottom solver in MLMG preconditioners by @WeiqunZhang in #7350
- Fix Ex in 1D collocated electrostatic solve by @WeiqunZhang in #7355
- Fix imaginary-part indices in RZ EvolveF by @WeiqunZhang in #7354
- Fix z coordinate in 1D FieldReduction by @WeiqunZhang in #7356
- Apply particle offset to BTD old-attribute pointers by @WeiqunZhang in #7361
- Fix bremsstrahlung multiplier reduction when probability is capped by @WeiqunZhang in #7358
- Fix upper Neumann row in 1D tridiagonal Poisson solve by @WeiqunZhang in #7363
- Fix edge direction for vacuum cells in 1D radial fluids by @WeiqunZhang in #7362
- Fix sign of imaginary part in RZ implicit Villasenor E gather by @WeiqunZhang in #7360
- Use coarse-patch spectral solver for G_cp by @WeiqunZhang in #7359
- Expose MLMG bottom solver and coarse level controls for the ES solvers by @roelof-groenewald in #7228
- Add sub-stepping with the ThetaImplicitEM solver by @dpgrote in #7000
- Darwin reduced diagnostic output fix by @roelof-groenewald in #7062
- Allocate B_old with the same number of components as Bfield_fp by @ax3l in #7371
- Do not skip injection sources after a single particle by @WeiqunZhang in #7369
- Do not scale dt per level in ApplyDtLimiters without subcycling by @WeiqunZhang in #7377
- Fix OpenMP race in FieldPoyntingFlux by @WeiqunZhang in #7378
- Use particle weight when Histogram2D has no value_function by @WeiqunZhang in #7370
- Read per-source q_tot, npart_real and focal_distance by @WeiqunZhang in #7379
- Fix radial Poisson BCs in RZ when rmin > 0 by @WeiqunZhang in #7375
- Fix reduced mass in inverse bremsstrahlung Coulomb log by @WeiqunZhang in #7374
- PICMI: multiply by density_scale instead of replacing by @WeiqunZhang in #7373
- Pydantic PICMI in WarpX by @ax3l in #6927
- Require warpx.read_fields_from_path for read_from_file by @WeiqunZhang in #7382
- Read int component names in BTD species header by @WeiqunZhang in #7381
- Damp guard cells once per box, not once per tile by @WeiqunZhang in #7376
- Add an option to load temperature in eV for the Maxwellian momentum distribution by @oshapoval in #7184
- Replace theta with temperature_in_eV for the Maxwell-Juttner momentum distribution by @oshapoval in #7205
- Fix PETSc SNES Nonlinear Solver by @debog in #7383
- Fixed formatting of injection sources documentation by @dpgrote in #7387
- Remove obsolete files by @EZoni in #7386
- Implicit solver: AlgMG preconditioner (pc_algmg) by @WeiqunZhang in #7349
- Fix substepping in SemiImplicitEM by @dpgrote in #7392
- Sum guard cells of J deposited for diagnostics by @WeiqunZhang in #7372
- Hybrid-PIC: fix external-field double-add on segmented stepping by @clarkse-he in #7161
- Add options to lag the PC update for nonlinear implicit solvers. by @JustinRayAngus in #7351
- Restore electric field before retrying semi-implicit substeps by @JustinRayAngus in #7398
- Docs: PETSc SNES nonlinear solver by @debog in #7389
- Fix rotation defaults in velocity coincidence thinning by @WeiqunZhang in #7408
- Use ionization level in Python deposit_charge by @WeiqunZhang in #7406
- WarpXSumGuardCells: always specify icomp and ncomp parameters by @JustinRayAngus in #7409
- Restore pc_jacobi guard and mass-matrix deposit skip for lagged PC updates by @debog in #7411
- Fix:
load_inputs_filewith Comment afterFILEby @ax3l in #7410 - Dependencies: weekly update by @github-actions[bot] in #7397
- Use NoGpuSync for MLMG in the curl-curl and Darwin preconditioners by @RemiLehe in #7388
- Nonlinear solvers: consolidate input parameters and shared code by @debog in #7412
- Identical Algorithm/Behavior between PETSc SNES and Native Newton Solver by @debog in #7384
- Darwin: use consistent B gather in predictor and corrector steps by @roelof-groenewald in #7315
- Hybrid-PIC electron energy equation: keep grad(P_e) in the B update, put the QDSMC markers on the grid nodes by @prkkumar-he in #7344
- Rework pulsed decay to a particle basis by @dpgrote in #7343
- CI: Re-enable MSVC build by @ax3l in #7131
- Fix:
load_inputs_filein Jupyter by @ax3l in #7282 - Pulsed decay minor fixes by @JustinRayAngus in #7419
- Bump github/codeql-action from 4.38.2 to 4.38.3 by @dependabot[bot] in #7414
- CI: Pin GitHub Actions to Commit SHAs by @ax3l in #7430
- Support wider mass-matrix stencils for the PC-only path by @JustinRayAngus in #7314
- Allow the PETSc LU preconditioner on a single GPU by @debog in #7428
- ParticleUtils.H : use amrex::Math::sincos(angle) instead of std::sin(angle)/std::cos(angle) by @lucafedeli88 in #7422
- C++20 : introduce std::ranges in TextMsg.cpp and MsgLogger.cpp by @lucafedeli88 in #7135
- Unify charge- and current-density finalization across geometries and solvers: Part 1 of 3 by @JustinRayAngus in #7341
- WarpXFluidContainer.cpp: add missing constexpr in some if clauses by @lucafedeli88 in #7437
- VirtualPhotonCreation.cpp: add checks for qed_virtual_photons_min_energy and qed_virtual_photons_multiplier by @lucafedeli88 in #6539
- Bump advanced-security/filter-sarif from 1.0.1 to 1.1 by @dependabot[bot] in #7431
- Fix segfault with mesh refinement + embedded boundary + electromagnetic solver by @RemiLehe in #6292
- Remove the Python function
deposit_charge_densityby @RemiLehe in #7439 - Implicit solver: expose and retune the pc_algmg Chebyshev smoother by @WeiqunZhang in #7423
- Dependencies: weekly update by @github-actions[bot] in #7449
- Fuse mass-matrix transfers into a single kernel per grid box by @JustinRayAngus in #7434
- Bump github/codeql-action/upload-sarif from 4.38.3 to 4.39.0 by @dependabot[bot] in #7455
- PETSc PC matrix: exact preallocation at the first assembly by @debog in #7460
- Dependabot: Group GitHub Actions Updates by @ax3l in #7462
- Move the sparse preconditioner matrix out of MatrixPC by @debog in #7424
- Python: default
finalize=Trueinget_charge_densityby @RemiLehe in #7451 - Fix EB potential parser mode when returning to time-only expressions by @n01r in #7420
- Use std::ranges::fill and std::ranges::for_each instead of std::ranges and std::for_each for compactness and clarity by @lucafedeli88 in #7452
- Implicit solvers: periodically re-initialize PETSc on GPU builds by @WeiqunZhang in #7453
- Record build times in CDash by @EZoni in #7405
- Bump the github-actions group with 3 updates by @dependabot[bot] in #7463
- Fields.H : isFieldArray is now a [[nodiscard]] constexpr function by @lucafedeli88 in #7416
- Fix
ArrayFieldTypesand thevector_potential_fp_nodalsolver vector by @ax3l in #7465 - Docs: document usage of /dev/null for warpx.used_inputs_file by @lucafedeli88 in #7429
- Unify charge- and current-density finalization across geometries and solvers: Part 2 of 3 by @JustinRayAngus in #7395
- Fix RSPHERE warnings flagged by CDash by @EZoni in #7472
- MatrixSolverOps interface; JacobiPC built on it by @debog in #7425
- Hybrid-PIC: preserve adaptive substeps on checkpoint restart by @tomzhu0225 in #7393
- Bump the github-actions group with 5 updates by @dependabot[bot] in #7478
- Dependencies: weekly update by @github-actions[bot] in #7475
- C++20 : use constraints by @lucafedeli88 in #6902
- Implicit solvers: skip PETSc re-init when there is no nonlinear solver by @roelof-groenewald in #7481
- Compile parse_field_function on a single thread by @PhysicsDan in #7485
- Bug fix in Darwin model - update E guard cells before predictor push by @roelof-groenewald in #7483
- Fewer guard cell exchanges when applying Darwin linear operator by @RemiLehe in #7177
- Darwin: Allow lower particle shape for MS than ES parts of the model by @roelof-groenewald in #7318
- Docs: Robust Doxygen Tag File Downloads by @ax3l in #7257
- Abort if particle tiling is turned on in a GPU build by @RemiLehe in #7489
- Move PC mass-matrix updates out of residual evaluation by @JustinRayAngus in #7436
- Checksums: print key of maximum relative error by @EZoni in #7498
- Unify charge- and current-density finalization across geometries and solvers: Part 3 of 3 by @JustinRayAngus in #7418
- Prohibit unphysical electromagnetic terms for 1D RSPHERE geometry by @JustinRayAngus in #7345
- Add input parameter warpx.tiny_profiler_flush_interval by @dpgrote in #6848
- Exchange RZ PML corner guard cells for all fields in MoveWindow by @WeiqunZhang in #7380
- Release: WarpX 26.11 by @github-actions[bot] in #7493
Full Changelog: 26.10...26.11