Merge devin hs2 #148
Open
Merge devin hs2 #148
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* Python: `warpx.multifab` legacy signature Keep the legacy signature of the fully qualified name a bit longer to avoid breakage. Throw warnings for users to migrate. * Update Docs * Simplify Example String No f-string needed here. * Old API: Do not warn yet * Clang-Tidy
…ECP-WarpX#5324) * un-multiply done * Update Source/ablastr/fields/PoissonSolver.H --------- Co-authored-by: Remi Lehe <remi.lehe@normalesup.org>
Keep ownership on the C++ side.
Fix the missing include for `MultiLevelRegister`, which includes the MultiLevel MF types.
Fix the missing include for `MultiLevelRegister`, which includes the MultiLevel MF types.
ECP-WarpX#5289) * Fixed a bug where centering coefficients used in Magentostatic solver were not being initialized when doing an energy-conserving field gather. Signed-off-by: S. Eric Clark <25495882+clarkse@users.noreply.github.com> * [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci * Pointer type passed into the magnetostatic solver for current is no longer smart, so removed the call to get() to return the raw pointer. Signed-off-by: S. Eric Clark <25495882+clarkse@users.noreply.github.com> --------- Signed-off-by: S. Eric Clark <25495882+clarkse@users.noreply.github.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* Remove m_cell_centered_data from multifab map * Register cell_centered_data MultiFab * Cleaning & Member Variable Convention * Shorten: fields as variable Prepares to remove `WarpX` class altogether from here. --------- Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
* `FieldType`: Top Doxygen Comment Add a general top level comment. * Doxygen: Expand `AMREX_ENUM` * `FieldType`: Doxygen Oneline Use a comment form that will still work after it gets pasted and squashed into a single line after macro expansion. * Doc: Include in Sphinx * RTD: Doxygen 1.9.1 to latest
* Skip Guard in IGF Solver (Revert ECP-WarpX#5284) This reverts the inclusion of the guard cells when copying rho. We could not yet determine why, but with this change the FFT 3D solver in ImpactX does not converge to the analytical solutions anymore. * Reset Checksum
…CP-WarpX#5226) * General moving-window transformations in boosted-frame simulations * Default speed of moving window to speed of boosted frame * Extend simulation volume enough so that particles don't exit * Include moving window speed in diag_lo and diag_hi transformations * Modify bounds so as to produce the same results as in previous versions --------- Co-authored-by: Remi Lehe <remi.lehe@normalesup.org>
* AMReX: Weekly Update * pyAMReX: Weekly Update
Move the boundary handler to become an optional argument, which otherwise defaults to Dirichlet conditions, e.g., in non-EB cases. This simplifies the ImpactX implementation and fixes a linker issue with CUDA for ImpactX.
updates: - [github.com/astral-sh/ruff-pre-commit: v0.6.7 → v0.6.8](astral-sh/ruff-pre-commit@v0.6.7...v0.6.8) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Do not default-advertise to enable heFFTe on Perlmutter and Lonestar yet. Introduce user-facing default for all HPC systems at a later point when it makes more sense.
The package `cuda-nvtx-dev` does not seem to be part of the `cuda` package. https://anaconda.org/nvidia/repo This fixes: ``` CMake Error at build/_deps/fetchedamrex-src/Tools/CMake/AMReXParallelBackends.cmake:71 (target_link_libraries): Target "amrex_3d" links to: CUDA::nvToolsExt but the target was not found. Possible reasons include: * There is a typo in the target name. * A find_package call is missing for an IMPORTED target. * An ALIAS target is missing. Call Stack (most recent call first): build/_deps/fetchedamrex-src/Src/CMakeLists.txt:40 (include) ```
* use plasma current rather than total current in `HybridPICSolveE` * remove logic to subtract J_ext from plasma current in `JdispFunctor` * add one ghost cell to the hybrid-pic external current since we interpolate to a nodal grid * Fix Doxygen Signed-off-by: roelof-groenewald <regroenewald@gmail.com>
…pX#5308) When `pre-commit` gets auto-updated, we typically see a PR from the branch `pre-commit-ci-update-config`. This branch is created directly in the main fork (unlike the individual branches that WarpX contributors create from their own forks) and all CI checks run twice: - once for the activity on the PR that `pre-commit` automatically opens (these CI checks are labeled "PR automated") - once for the activity on the branch `pre-commit-ci-update-config` (these CI checks are labeled "individual CI") Here's an example:  On top of this, once the PR is merged, CI runs a third time, because the merge is counted as activity on the branch `development` (again "individual CI"). We should be able to safely skip "individual CI" for the activity on the branch `pre-commit-ci-update-config`. This PR should do the trick, although it's good to double check the syntax for GitHub Actions and Azure pipelines. My understanding is that the cleanup-cache and post-PR workflows don't need to be updated, also to be double checked. --------- Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
Added the ability to synchronize particle velocities and positions inside a python script via simulation.extension.warx.synchronize(). Initiated after discussion with @roelof-groenewald in ECP-WarpX#5331. Implemented and tested by comparing to Turner's benchmark results (https://doi.org/10.1063/1.4775084). <p align="center"> <img src="https://github.com/user-attachments/assets/366aab43-359d-440c-a67c-7eef97028f4f">
This feature replaces the InitializeExternalFieldsOnGridUsingParser with a more general function ComputeExternalFieldOnGridUsingParser. The new function takes parsed functions with four dimensions (x,y,z,t), so that it can be used to evaluate functions throughout the simulation, for example time-dependent external fields. The function ComputeExternalFieldOnGridUsingParser has optional edge length and face areas. --------- Co-authored-by: Kristoffer Lindvall <kristoffer.lindvall@novatronfusion.com>
Prepare the October release of WarpX: ```bash # update dependencies ./Tools/Release/updateAMReX.py ./Tools/Release/updatePICSAR.py # no changes, still 24.09 ./Tools/Release/updatepyAMReX.py # bump version number ./Tools/Release/newVersion.sh ``` Following this workflow: https://warpx.readthedocs.io/en/latest/maintenance/release.html
* Explicitly use gcc-12, because gcc points to gcc/7.5. * Load the non-cray version of hdf5 because the cray version has some cmake issues.
The script was not robust enough for the black/ruff formatting updates we applied. This affected the 24.09 and 24.10 release tags. This fixes it.
Update installation instructions for the Lawrencium Cluster (LBNL).
Having a consistent set of base input files can make default testing easier for package managers (e.g., Spack, Conda). Example pointed out by @ax3l: https://github.com/spack/spack/blob/d21577803f7acb4cc1a5b8144762ea052f67ecab/var/spack/repos/builtin/packages/warpx/package.py#L469
Now that all MultiFabs are registered in the new system, the old `multifab_map` can be removed. Note that the `imultifab_map` is still needed since the new system does not yet work with `iMultiFabs`. This breaks the Python API since it removes `warpx.field` which was returning the old map.
Fix ECP-WarpX#5206 and prepare for ECP-WarpX#5256: - [x] Add missing checksums - [x] `test_1d_background_mcc_picmi` - [x] `test_2d_background_mcc_picmi` - [x] `test_3d_particle_absorption` - [x] `test_2d_dirichlet_bc` - [x] `test_2d_dirichlet_bc_picmi` - [x] `test_2d_field_probe` - [x] `test_3d_embedded_boundary_picmi` - [x] `test_2d_particle_reflection_picmi` - [x] `test_rz_laser_acceleration_opmd` - [x] `test_2d_runtime_components_picmi` - [x] Clean up calls to `evaluate_checksum` Calls to `evaluate_checksum` are now uniform across the analysis scripts, making it easier to automatize further changes. Hopefully the uniformity will be preserved by copy/pasting the necessary code: - ```py import os import sys ``` - ```py sys.path.insert(1, "../../../../warpx/Regression/Checksum/") from checksumAPI import evaluate_checksum ``` - ```py # compare checksums evaluate_checksum( test_name=os.path.split(os.getcwd())[1], output_file=sys.argv[1], ) ``` with this last one possibly modified by adding other arguments, e.g., `output_format`, `rtol`, `do_particles`, etc.
Fix some RST `verbatim`, which needs two backticks compared to Markdown.
**Magnetic Reconnection: An Alternative Explanation of Radio Emission in Galaxy Clusters** by Subham Ghosh and Pallavi Bhat was just published. https://10.3847/2041-8213/ad9f2d --------- Co-authored-by: Edoardo Zoni <59625522+EZoni@users.noreply.github.com>
Co-authored-by: Remi Lehe <remi.lehe@normalesup.org>
…rpX#5619) The calculation of Fornberg stencil coefficients is rather general, and it can be shared with other projects of the BLAST family. Therefore, this PR moves the responsible functions into `ablastr`. Specifically, the PR does the following: - 2 new files (`FiniteDifference.H` and `FiniteDifference.cpp`) are created under `ablastr/math` (`CMakeLists.txt` and `Make.package` accordingly) - the static method of the WarpX class `getFornbergStencilCoefficients` and the `ReorderFornbergCoefficients` function (originally defined in an anonymous namespace in `WarpX.cpp`) are moved to these new files, inside the namespace `ablastr::math` - the two methods are minimally adapted (e.g., `AMREX_ALWAYS_ASSERT_WITH_MESSAGE` becomes `ABLASTR_ALWAYS_ASSERT_WITH_MESSAGE`) - `WarpX.cpp` and `SpectralKSpace.cpp` (where the aforementioned functions were called) are updated Note that with this PR `SpectralKSpace.cpp` does not need anymore to include the heavy `WarpX.H` header.
<!--pre-commit.ci start--> updates: - [github.com/astral-sh/ruff-pre-commit: v0.9.4 → v0.9.6](astral-sh/ruff-pre-commit@v0.9.4...v0.9.6) <!--pre-commit.ci end--> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
…#5642) This PR turns the static variable `em_solver_medium` of the WarpX class into a private non-static member variable : `m_em_solver_medium` . This is done with the aim of reducing the usage of static variables in WarpX.
This PR bumps the version used for `clang-tidy` CI tests from 16 to 17. It also addresses all the issues found with the upgraded tool. To be merged **after** ECP-WarpX#5592 ✅ ### The issues found 🧐 and fixed 🛠️ with the upgraded tool are the following : - [bugprone-switch-missing-default-case](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/bugprone/switch-missing-default-case.html) A newly introduced check to flag `switch` statements without a `default` case (unless the argument is an `enum`) - [cppcoreguidelines-rvalue-reference-param-not-moved](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/cppcoreguidelines/rvalue-reference-param-not-moved.html) A newly introduced check to flag when an rvalue reference argument of a function is never moved inside the function body.⚠️ **Warning**: in order to have this check compatible with [performance-move-const-arg](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/performance/move-const-arg.html) I had to set `performance-move-const-arg.CheckTriviallyCopyableMove` to `false` (specifically for the three methods in `ablastr::utils::msg_logger` accepting `std::vector<char>::const_iterator&& rit` arguments). - [misc-header-include-cycle](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/misc/header-include-cycle.html) A newly introduced check to prevent cyclic header inclusions. - [modernize-type-traits](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/modernize/type-traits.html) A newly introduced check. The idea is to replace currencies of, e.g., `std::is_integral<T>::value`, with the less verbose alternative `std::is_integral_v<T>` - [performance-avoid-endl](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/performance/avoid-endl.html) A newly introduced check. The idea is to replace `<< std::endl` with `\n`, since `endl` also forces a flush of the stream. In few cases flushing the buffer is actually the desired behavior. Typically, this happens when we want to write to `std::cerr`, which is however automatically flushed after each write operation. In cases where actually flushing to `std::cout` is the desired behavior one can do `<< \n << std::flush `, which is arguably more transparent than `<< std::endl`. - [performance-noexcept-swap](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/performance/noexcept-swap.html) For performance reasons it is better if `swap` functions are declared as `noexcept`, in order to allow the compiler to perform more aggressive optimizations. In any case, we can use the AMReX function `amrex::Swap`, which is `noexcept`. ### 🔄 Re-enabled checks: - [readability-misleading-indentation](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/readability/misleading-indentation.html) This check was already available in v16, but a bug led to false positives. The bug has been corrected in v17 of the tool, so we can re-enable the check. ### ⛔ The PR excludes the following checks : - [cppcoreguidelines-missing-std-forward](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/cppcoreguidelines/missing-std-forward.html) A newly introduced check that warns when a forwarding reference parameter is not forwarded. In order to comply with this check I think that I have to pass some parameters by reference to lambda functions inside `ParallelFor` constructs. However, this leads to issues when we compile for GPUs. Therefore, I think that the best solution is to exclude this check. See an example below (for `PredFunc&& filter` ): ``` amrex::ParallelForRNG(np, [=,&filter] AMREX_GPU_DEVICE (int i, amrex::RandomEngine const& engine) { p_mask[i] = filter(src_data, i, engine); }); ``` - [misc-include-cleaner](https://releases.llvm.org/17.0.1/tools/clang/tools/extra/docs/clang-tidy/checks/misc/include-cleaner.html) It would be awesome to include this check. However, as it is now implemented, it has no notion of "associated headers". For instance, let's suppose that the header `MyClass.H` has `#include<string>` and that `MyClass.cpp` has `#include "MyClass.H"` and uses `std:string` somewhere. In this case, the check raises a warning stating that you should include `<string>` in `MyClass.cpp` even if it is transitively included via the associate header `MyClass.H` . For this reason, for the moment, it is better to periodically check headers with the `IWYU` tool.
…CP-WarpX#5579) This PR moves the initialization of the warning manager from the very large `ReadParameters` function of the WarpX class to a free function inside `WarpXInit.H/cpp` . This function is now called by the constructor of the WarpX class. The final goal is to simplify the WarpX class.
….cpp (ECP-WarpX#5609) This PR moves the static function `shiftMF` from the WarpX class to an anonymous namespace in `WarpXMovingWindow.cpp`, where it is actually used. This is done to simplify the Warpx class.
Weekly update to latest AMReX. Weekly update to latest pyAMReX. Weekly update to latest PICSAR (no changes). ```console ./Tools/Release/updateAMReX.py ./Tools/Release/updatepyAMReX.py ./Tools/Release/updatePICSAR.py ``` --------- Signed-off-by: Axel Huebl <axel.huebl@plasma.ninja> Co-authored-by: Weiqun Zhang <weiqunzhang@lbl.gov>
This PR transforms the WarpX member functions `MarkReducedShapeCells`, `MarkUpdateCellsStairCase`, `MarkUpdateECellsECT`, `MarkUpdateBCellsECT`, `MarkExtensionCells` into pure functions inside the namespace `warpx::embedded_boundary`, together with `ComputeEdgeLengths`, `ComputeFaceAreas`, `ScaleEdges`, and `ScaleAreas`. The source files containing these functions are renamed as `EmbeddedBoundaryInit.H/cpp` , since these functions are called only during the initialization.
…to&` (ECP-WarpX#5656) This PR simplifies the return type of a method of the WarpX class by replacing: ``` # ifdef WARPX_DIM_RZ SpectralSolverRZ& # else SpectralSolver& # endif ``` with ``` auto& ```
Adds a luminosity diagnostic differentiated in the energies of two
colliding species, called `DifferentialLuminosity2D`.
It is defined as follows:
```math
\begin{align*}
\frac{d^2\mathcal{L}}{dE_1 dE_2}(E_1, E_2, t) = \int_0^t dt'\int d\boldsymbol{x}\, & \int d\boldsymbol{p}_1 \int d\boldsymbol{p}_2\; \sqrt{ |\boldsymbol{v}_1 - \boldsymbol{v}_2|^2 - |\boldsymbol{v}_1\times\boldsymbol{v}_2|^2/c^2} \\
& f_1(\boldsymbol{x}, \boldsymbol{p}_1, t')f_2(\boldsymbol{x}, \boldsymbol{p}_2, t') \delta(E_1 - E_1(\boldsymbol{p}_1) \delta( E_2 - E_2(\boldsymbol{p}_2))
\end{align*}
```
where:
* $\boldsymbol{p}_i$ is the momentum of a particle of species $i$
* $E_i$ is the energy of a particle of species $i$, $E_i
(\boldsymbol{p}_i) = \sqrt{m_1^2c^4 + c^2 |\boldsymbol{p}_i|^2}$
* $f_i$ is the distribution function of species $i$, normalized such
that $\int \int f(\boldsymbol{x} \boldsymbol{p}, t )d\boldsymbol{x}
d\boldsymbol{p} = N$, the number of particles in species $i$ at time $t$
The 2D differential luminosity is given in units of $\text{m}^{-2} \
\text{eV}^{-2}$.
The user must specify the minimum, maximum, and number of bins to
discretize the $E_1$ and $E_2$ axes.
The computation of this diagnostic is similar to that of
`ParticleHistogram2D`.
The output is a folder containing a set of openPMD files.
The values of the diagnostic are stored in a record labeled
`d2L_dE1_dE2`, with axes `E1` and `E2`.
---------
Co-authored-by: Remi Lehe <remi.lehe@normalesup.org>
…-WarpX#5644) `SetDotMask`, a member function of the WarpX class, is only used inside the member function `getFieldDotMaskPointer` . This PR turns it into a pure function and moves it into an anonymous namespace inside `WarpX.cpp`. This (slightly) simplifies the WarpX class header.
Update to latest AMReX version to pull the latest bug fix in AMReX-Codes/amrex#4333.
This PR allows for the addition of external fields through the particle fields analytical interface. This is useful for field splitting external vs. self fields in the hybrid ohm's law solver. --------- Signed-off-by: S. Eric Clark <25495882+clarkse@users.noreply.github.com> Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Roelof Groenewald <40245517+roelof-groenewald@users.noreply.github.com>
This is a temporary fix, setting a boundary condition for the Curl-Curl preconditioner for the implicit solver. This now sets the BC to Dirichlet for the PEC regions. A correct solution would have to be implemented in the curl-curl solver because of the split between the PEC and insulator sections.
- [x] profile: avoid repetition, use `SW_DIR` variable as in `install_...` scripts - [x] move from CFS to PSCRATCH (more stable, faster, where the binary lives); uses an undocumented, purge-exempt location for container images/software - [x] build our own boost (SW stack consistency) - [x] get our own CCache (prior one is gone) - [x] RT tested
As said in the title, this is just a small PR to make the `HybridPICSolveE` kernels more readable. --------- Signed-off-by: roelof-groenewald <regroenewald@gmail.com>
…ev) and doQEDEvents with doQEDEvents(lev) (ECP-WarpX#5671) `doFieldIonization(lev) ` is called only once, inside `doFieldIonization` , which is simply a loop over the levels calling for each level `doFieldIonization(lev) `. The same happens for `doQEDEvents`. In order to simplify the interface of the WarpX class, I would like to propose to drop `doFieldIonization(lev) ` and `doQEDEvents(lev) `, and to integrate their code respectively in `doFieldIonization` and `doQEDEvents`.
…arpX#5670) `AverageAndPackFields` and `prepareFields` are not implemented. Therefore, this PR removes their declaration from the WarpX header.
The box metadata used in this script follows the AMReX conventions. We want "zyx" in 3D and "yx" in 2D.
…tData.cpp (ECP-WarpX#5658) `PrintDtDxDyDz` is used only twice in `WarpXInitData.cpp`. Therefore, this PR turns it from a method of the WarpX class to a simple function inside an anonymous namespace in `WarpXInitData.cpp`
Weekly update to latest AMReX. Weekly update to latest pyAMReX (no changes). Weekly update to latest PICSAR (no changes). ```console ./Tools/Release/updateAMReX.py ./Tools/Release/updatepyAMReX.py ./Tools/Release/updatePICSAR.py ``` This pulls in AMReX-Codes/amrex#4337, fixing regressions from ECP-WarpX#5669 (GPU segfaults on particle redistribute) Signed-off-by: Axel Huebl <axel.huebl@plasma.ninja>
…-WarpX#5685) `NodalSyncJ` and `NodalSyncRho` are member functions of the WarpX class, but they are never used. Therefore, this PR removes them.
Changes in Perlmutter GPU job script: from `#SBATCH --cpus-per-task=16` to `#SBATCH --cpus-per-task=32`. This is to request (v)cores in consecutive blocks. GPU 3 is closest to CPU cores 0-15, 64-79, GPU 2 to CPU cores 16-31, 80-95, ... If `--cpus-per-task=16`, MPI ranks 0 and 1 are mapped to cores 0 and 8. If `--cpus-per-task=32`, MPI ranks 0 and 1 are mapped to cores 0 and 16. Visual representation  --------- Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
New PoP by Ashwyn Sam et al.: https://doi.org/10.1063/5.0249525
… MergeDevinHS2
RevathiJambunathan
force-pushed
the
HarrisSheetinX
branch
from
019ea2c to
aa536a0
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.