Age | Commit message (Collapse) | Author | Files | Lines |
|
* Spack Envs: Debian
Add default compilers on Debian to our Spack environments.
* Debian: have `lsb-release` installed
See: https://github.com/spack/spack/pull/32479
|
|
When warpx.refine_plasma=1, there was an error in the logic like below.
if (lrefine_injection) {
Box fine_overlap_box = ...;
if (fine_overlap_box.ok()) {
pcounts[index] = ...;
} // the else part was missing
} else {
pcounts[index] = ...;
}
|
|
Add nodal field gather functions to ABLASTR for scalar and vector
fields. This currently only implements 1st order (CIC) shape.
Co-authored-by: Remi Lehe <rlehe@lbl.gov>
Co-authored-by: Remi Lehe <rlehe@lbl.gov>
|
|
|
|
* Perlmutter: Work-Around CUDA-Aware MPI & Slurm
There are known HPE bugs on Perlmutter that can blow up
simulations (segfault) with CUDA-aware MPI.
We avoid the respective Slurm options now and just manually
control the exposed GPUs per MPI rank.
* Add: `gpus-per-node`
|
|
* fix few issues found with clang tidy
* fix bug
|
|
|
|
* Python: 3.7+
Python 3.6 is now end-of-life. Bump support to 3.7+.
* CUDA CI: Python3 Update
|
|
* add mention of callbacks and some libwarpx functions to the docs
* reformatted various function docstrings in `_libwarpx.py` and added them to the docs
* Add subsection
Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
|
|
* Fix Bugs w/ Current Correction and Vay Deposition
* Vay Deposition and Current Correction Cannot be Combined Together
* Add Comment for Future Implementation of Vay Deposition w/ MR
* Add Comment for Future Implementation of Vay Deposition w/ MR
* Define SyncCurrentAndRho, Clean Up
* Vay Deposition: Remove Extra FFT of Rho
* Fix Bug in RZ Geometry (Double Filtering)
* Add 2D Galilean Test w/o Periodic Single Box
* Add RZ Galilean Test w/o Periodic Single Box
* Add 3D Galilean Test w/o Periodic Single Box
|
|
* Fixed plasma lens residence correction to allow a short lens
* Added CI test with a short plasma lens
|
|
* order-independent rho_<species>
* add test to check rho_<species> is order-agnostic
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* make analysis naming more consistent
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
|
|
* Use proper units for momenta, in BoundaryScrapingDiagnostic
* Fix fetch of attribs in particlesConvertUnits
Co-authored-by: Dave Grote <grote1@llnl.gov>
|
|
* avoid details bugs with nsight system analysis
I added the header of the sbatch file for nsight system analysis, a little description of the useful output files for the analysis and a warning for the execution of the sbatch file.
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Update Docs/source/developers/profiling.rst
Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
* Update Docs/source/developers/profiling.rst
Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
* Update profiling.rst
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
* Fix: remove ill-placed anchor
* Fix ill-removed section
* Fix Formatting
* Update Wording
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
|
|
|
|
- Call `setVal` with default arguments
(number of components and ghost cells)
- Affects only functions used with multi-J time stepping:
- `MultiParticleContainer::DepositCurrent`
- `MultiParticleContainer::DepositCharge`
|
|
Print a performance hint if too many boxes are used per GPU.
|
|
Add Spack environment files for developer desktop/laptop computers.
|
|
updates:
- [github.com/Lucas-C/pre-commit-hooks: v1.3.0 → v1.3.1](https://github.com/Lucas-C/pre-commit-hooks/compare/v1.3.0...v1.3.1)
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
|
|
|
|
* CI: Clean up Analysis of Galilean Tests
* Update Reference Energy Values
* Update Reference Energy Tolerances
* Improve Inline Comment
|
|
Update comments in example input files: moving window now available along x, y, z.
|
|
|
|
* shift BTD lab frame bounds by half a cell to prevent round off errors
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
|
|
Add our latest NAPAC'22 paper to the documented references.
Will replace with a JACOW reference once proceedings are out.
|
|
|
|
* Wheel: Separate Directory
The `pip_install` target did not yet work on Windows, because it has
no universal wildcard `*` support for our pip install logic.
Fix this by creating the wheel in a sub-directory and installing it
by "finding all wheels for `pywarpx` in a given prefix".
- Avoid wildcards by using a separate directory.
- Avoid downloading deps during `wheel` build again.
* Fix: Manifest
follow-up to #3265
|
|
Workaround a OpenMPI bug that uses file locking on Lassen (OLCF).
Performance is still not going up...
|
|
* Allow filtering in z with RZ FDTD
* Update the error message about filtering in RZ
* Turned z filtering on in laser_acceleration/inputs_rz
* Update LaserAccelerationRZ benchmark, including z filtering
|
|
Add our latest NAPAC'22 paper to the documented references.
Will replace with a JACOW reference once proceedings are out.
|
|
See if this missing include, which would import dllexport, fixes
symbol errors on Windows in ImpactX.
|
|
Our Windows CI does not abort on the first failing command. This
could, in the future, lead to false-positive tests.
|
|
* Rearrange in preparation for lifting do_not_deposit higher in the call stack
* Make do_not_deposit public, so callers can test on it
* Make all DepositCharge call sites respect WarpXParticleContainer::do_not_deposit
* Drop check for do_not_deposit in DepositCharge
* Fix comment to accurately describe np_to_depose
* Fix logic to account for other logic in AddSpaceChargeField
* Match zero charge density array to filled array in RZ PSATD case
* Structure control flow per @dpgrote's preference
* Reword comment for clarity
|
|
|
|
* Fill Guard Cells of Fields and Currents Independently
* Fill Guard Cells of Averaged Fields as Regular Fields
|
|
* Enable particle RZ BackTransformed diagnostics
* Add warning + compute theta
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
|
|
|
|
* `isort`: AMReX, WarpX, etc. as First Party
Mark the `amrex` and `pywarpx`, `picmistandard`, ... imports as first
party, so that they do not change if run locally or remotely pre/post
install.
|
|
* Initialize particle runtime attributes before calling AddNParticles
* Avoid Multiplication result converted to larger type warning
* Directly initialize runtime attributes inside AddNParticles
* Update doxygen comment
|
|
|
|
(#3294)
* Fix out of boundary access with Particle Scraping + Continuous Injection
* Explicit type of particles
* Change type of particles
|
|
|
|
This adds all currently defined PICMI parameters to the user-facing
docs.
|
|
* uise extended box array in setting par GDB for BTD
* add synchronize in BTD particle functor
|
|
|
|
* Checksum: Ignore cpu & id
Both are dependent on MPI parallelism and differ if run with
one or more MPI ranks, as well if run on CPU (NOACC, OMP) or
GPU (CUDA, HIP, SYCL).
* Benchmarks: Remove cpu/id from JSON
* Update Script Further
- fixes: field/species property, not species itself
- comments
Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja>
|
|
* Species variables in diagnostics: deal with runtime components
* Remove ParticleStringNames namespace
* Use const reference in iteration variable
* Fix ion acc 2D test
* Output extra (e.g. initialized later with python) comps by default
* Always output theta in RZ
* Fix QED tests
* Fix Laser Acceleration RZ test
|
|
We do `tolerance` and `abs_tolerance` checks via our checksums. This
avoids that we need to build the AMReX plotfile tools in CI, which:
- take long time to build
- use GNUmake
- fail to build on macOS M1
Thus, do not add these sections to CI tests.
|
|
* Docs: Update Intro PFS OLCF
1. put more emphasis on reading the system docs for HPC systems
2. duplicate information form the system docs about purge times,
because you already skipped 1. again, didn't you
* All Systems: BOLD read the Sys Docs
|
|
* Docs: Lassen HDF Perf. Hints
Copy the IBM GPFS Performance hints we use on Summit to Lassen.
* Remove duplicate line
|