Wakis v0.6.0

🚀 New Features

• 🖼️ Plotting

  • Unified plotting tools for both MPI and non-MPI simulations.
  • plot1D now supports field visualization independently of MPI use.
  • Plot2D supports parallel execution.
  • Error handling added for plot3D and plot3DonSTL when use_mpi = True.
  • Support for dpi and return_handles in plot utilities to further customize plots.
  • Enhanced examples and notebook suite visualization cells.

• 🧱 GridFIT3D

  • Added mpi_initialize() to handle domain decomposition (Z-slicing).
  • New method mpi_gather_asGrid() to retrieve the full global grid from distributed subdomains.
  • Full support for multi-GPU domain decomposition through cupy (CUDA-aware, Linux only).
  • Improved communication layer for subdomain synchronization via ghost cells.

• ⚡ SolverFIT3D

  • MPI-compatible time-stepping routine mpi_one_step() using a leapfrog scheme.
  • mpi_communicate() to send/receive boundary field values between subdomains.
  • mpi_gather() to retrieve field data as a NumPy array and mpi_gather_asField() to reconstruct a Field object.
  • MPI-safe support integrated into update() and field getter logic.
  • Introduced save_state() method for checkpointing during MPI runs together with load_state(). Added support for MPI save state too.
  • Feature in progress: multiGPU support (use_GPU=True when use_MPI=True) for distributed simulations (Linux only!).
  • Added missing docstrings (Numpy-style)

• 📥 Sources

  • Beam:
    • Added plot(t) to visualize beam current evolution.
    • Generalized update() to work with or without MPI.
    • Enhanced support for time-aligned injection with beta and MPI shifts.
    • New example for MPI+GPU simulation (topic in progress).

• 🌊 WakeSolver

  • Refactored to internally store the full longitudinal domain.
  • skip_cells now acts only at analysis level, preserving resolution.
    • add_space and use_edt retained for compatibility, but add_space is deprecated for new parameter skip_cells, more adequate to its utility (i.e. skip cells in the integration path).
  • Future-ready structure for distributed wake solving with MPI-aware GPU.
  • Improved numerical robustness by preventing indexing errors in WakePotential integration -> solves issue #12
  • Enhanced extrapolation method with iddefix.
  • New example 004 for wakefield simulation with MPI+GPU configuration.

• 🛡️ Security & Documentation

  • Added SECURITY.md to describe supported versions and vulnerability reporting.
  • Improved installation guide with Miniforge (supports both Windows/Linux) and MPI setup instructions.
  • Added new issue templates for Bug Report and Feature Request with markdown formatting and emojis for readability.
  • Addition of the Physics Guide, with detailed physics models and numerical methods explanations.
  • User's guide updated to include Wake extrapolation with iddefix, Wake function calculation with neffint, and power loss calculation with BIHC.
  • Added a Table of Contents (ToC) to the documentation for easier navigation.
  • Expanded installation guide with multiGPU configuration and MPI-aware domain partitioning.

💗 Other Tag highlights

• 🔁 Nightly tests with GitHub Actions:

  • Enabled infrastructure for MPI-based test cases (test_003, test_005).
  • Improved test coverage for MPI and GPU simulations.
  • MultiGPU end-to-end tests for distributed domain synchronization.

• 📁 Examples:

  • notebook 003 → MPI wakefield simulation using mpi4py.
  • notebook 005 → Full MPI simulation inside Jupyter using ipyparallel + mpi4py.
  • New example 003 with MPI + GPU configuration for large-scale simulations.

• 📁 Notebooks:

  • 005 → Full MPI simulation inside Jupyter using ipyparallel + mpi4py.
  • New Jupyter notebook showcasing multiGPU configuration.

🐛 Bugfixes

• Fixed crash in plot3D and plot3DonSTL when use_mpi=True.
• Fixed default use_mpi=True to now default to False for general usage.
• Fixed a typo in beam injection routine solver.z.min().
• Fixed potential rounding error in wake potential integration with negligible performance impact (~0.1ns) -> solves issue #12
• Corrected default beam injection time to align with CST Wakefield Solver reference in beta<1 cases.
• Fixed minor typos in the documentation.
• Fixed synchronization issues with MPI runs when saving states.
• Resolved encoding issues when installing in Windows editable mode.
• Corrected result folder naming in GPU example 002.

👋👩‍💻 New Contributors

• @mctfr – PR #11 Contributed improvements to installation instructions in the documentation.

📝 Full changelog

git log v0.5.1... --date=short --pretty=format:"* %ad %d %s (%aN)" | copy

| 78 commits | 📚 Docs | 🧪 Tests | 🐛 Fixes | 🎨 Style | ✨ Features | Other | |----------------|---------|----------|-----------|------------|--------------| ----- | | % of Commits | 30.3% | 10.5% | 9.2% | 10.5% | 22.4% | 17.1 |

• 2025-05-27 docs: add docstrings for mpi_gather and mpi_gather_asField (elenafuengar)
• 2025-05-26 tests: fix mpi folder creation issues (elenafuengar)
• 2025-05-23 tests: only allowed 2 procs per CI workflow... (elenafuengar)
• 2025-05-23 tests: reduce MPI pipeline to only 007 (elenafuengar)
• 2025-05-23 fix: update run workflow name (elenafuengar)
• 2025-05-23 tests: add MPI CI/CD on manual py3.10 (elenafuengar)
• 2025-05-23 test: add 007 to pytest pipeline (elenafuengar)
• 2025-05-22 style: remove duplicates, prepare for release (elenafuengar)
• 2025-05-23 tests: passing 007 for single core and mpiexec! (elenafuengar)
• 2025-05-19 fix: nightly test report fix (elenafuengar)
• 2025-05-14 style: move mpl.plt import (elenafuengar)
• 2025-05-13 docs: update release draft, prepare for deployment (elenafuengar)
• 2025-05-09 feature: multiGPU working (Ubuntu) -but needs optimization (elenafuengar)
• 2025-05-09 style: revise GPU example 002, fix folder result name (Elena De La Fuente Garcia)
• 2025-05-09 build: fix encoding when installing editable in Windows (Elena De La Fuente Garcia)
• 2025-05-08 docs: add miniforge (Windows/Linux) to python installation guide (elenafuengar)
• 2025-05-08 docs: fix reference typo (elenafuengar)
• 2025-05-07 feature: support save_state for MPI runs (elenafuengar)
• 2025-05-06 docs: fix few mistakes spotted after RTD deployment (elenafuengar)
• 2025-05-06 docs: minor changes to adapt to the physics guide content (elenafuengar)
• 2025-05-06 docs: revised physics guide, make html passed (elenafuengar)
• 2025-05-05 docs: first version of physics guide (elenafuengar)
• 2025-05-02 feat: new example for MPI+GPU simulation (in progress) (elenafuengar)
• 2025-04-29 docs: update installation with MPI-GPU findings (elenafuengar)
• 2025-04-24 Update README.md (Elena de la Fuente García)
• 2025-04-24 Create SECURITY.md (Elena de la Fuente García)
• 2025-04-24 Update and rename feature-request-💡.md to feature-request.md (Elena de la Fuente García)
• 2025-04-24 Update bug-report.md (Elena de la Fuente García)
• 2025-04-24 Update and rename bug-report-🐛.md to bug-report.md (Elena de la Fuente García)
• 2025-04-24 Update and rename 💡feature-request-.md to feature-request-💡.md (Elena de la Fuente García)
• 2025-04-24 Add issue templates (bug & feature) (Elena de la Fuente García)
• 2025-04-09 docs: small typo in readme (Elena de la Fuente García)
• 2025-04-07 docs: update readme with playground contents (elenafuengar)
• 2025-04-02 build: update release draft version to 0.6.0 (elenafuengar)
• 2025-04-02 tests: add MPI test files and test script in progress (elenafuengar)
• 2025-04-02 build: add neffint, iddefix and bihc as dependencies (elenafuengar)
• 2025-04-02 style: fix results folder and plot kwargs (elenafuengar)
• 2025-03-29 docs: update release.md (elenafuengar)
• 2025-03-29 docs: small typo (elenafuengar)
• 2025-03-27 style: revised notebook 004, in particular the iddefix extrapolation (elenafuengar)
• 2025-03-27 feature: include wakefield simulation in example 003 (elenafuengar)
• 2025-03-27 docs: fix sidebar TOC in conf.py, add TOC to installation and user guide, minor fixes (elenafuengar)
• 2025-03-27 feature: add wakefield simulation to MPI example and extrapolation to fully decayed (elenafuengar)
• 2025-03-27 feature: MPI wakefield simulation with solver.wakesolve is now working (elenafuengar)
• 2025-03-27 docs: update MPI installation after testing on imp machines (elenafuengar)
• 2025-03-27 docs: fix in index.md for missing logo (elenafuengar)
• 2025-03-26 docs: prepare v0.5.2 release (elenafuengar)
• 2025-03-26 feature: support for MPI in wakesolve in progress + refact: wakesolve saves now all longitudinal values, skip_cells will only be applied inside WakeSolver. add_space and use_edt kept for legacy (elenafuengar)
• 2025-03-26 test: adjust tes_001 to WakeSolver refactor in previous commit (elenafuengar)
• 2025-03-26 fix: add if statement in wake potential calculation to catch rounding errors (profiler indicates only 0.1ns overhead) + refact: add_space now deprecated for skip_cells'. This parameter now adjusts the slicing of z instead of modyfing zmin and zmax, since solver.wakesolve` will now save all the longitudinal data (elenafuengar)
• 2025-03-26 fix: typo in solver.z.min() (elenafuengar)
• 2025-03-25 fix: change default to False for MPI (elenafuengar)
• 2025-03-21 feature: notebook 005 containing working MPI example using ipyparallel and mpi4py, using MPI methods inside Grid and Solver. Beautiful <3 (elenafuengar)
• 2025-03-21 feature: working MPI script for an electromagnetic simulation, using new MPI methods inside Grid and Solver (elenafuengar)
• 2025-03-21 style: change BCs for MPI subdmains to mpi for code readability (elenafuengar)
• 2025-03-21 feature: MPI implementation in SolverFIT3D, with mpi_one_step() to perform MPI leapfrog update, mpi_communicate() to send/recv information between mpisubdomains, mpi_gather() to retrieve the global field as a np.array for the specific x,y,z and componen, and mpi_gather_asField() to retrieve the global field as a Field instance (elenafuengar)
• 2025-03-21 feature: MPI implementation in GridFIT3D inside init, with mpi_initialize() to generate the z-subdomains and mpi_gather_asGrid() to retrieve the global grid (elenafuengar)
• 2025-03-21 fix: error handling for plot3D and plot3DonSTL when use_mpi = True -not supported (elenafuengar)
• 2025-03-21 feature: support for MPI and non-MPI inside plot1D -input script agnostic! (elenafuengar)
• 2025-03-21 feature: support for MPI and non MPI plots in Plot2D (elenafuengar)
• 2025-03-21 refactor: include MPI support inside uptade() (elenafuengar)
• 2025-03-21 style: remove check for component Abs, using getitem supported key (elenafuengar)
• 2025-03-20 feature: support key[3] = 'abs' or 'Abs' in getitem (elenafuengar)
• 2025-03-20 feature: update script to use MPI functions inside solver (elenafuengar)
• 2025-03-20 feature: include MPI functions inside solver to run MPI simulations with mpi4py and openmpi (elenafuengar)
• 2025-03-19 refact: change MPI update to use global ZMIN when solver hassattr, add plot(t) function (elenafuengar)
• 2025-03-18 Merge pull request #11 from mctfr/patch-1 (Elena de la Fuente García)
• 2025-03-18 Update installation.md (Manuel Cotelo Ferreiro)
• 2025-03-18 docs: revise docstrings (elenafuengar)
• 2025-03-18 docs: major revision of user guide, update index and installation guide to include MPI setup (elenafuengar)
• 2025-03-16 docs: update README & citation (elenafuengar)
• 2025-03-14 docs: add PR template (elenafuengar)
• 2025-03-13 fix: update default injection time to account for beta (elenafuengar)
• 2025-03-13 feature: add dpi and return_handles as kwargs (elenafuengar)

Full Changelog: https://github.com/ImpedanCEI/wakis/compare/v0.5.1...v0.6.0