- Sep 29, 2022
-
-
Johnson, Seth R. authored
* Add release docs for v0.1.2 * Add release notes and update documentation * Fix issues with code documentation * Remove comments * Add release process, contribution changes, dev guide tweaks * Incorporate review feedback * Update readme and ORNL TM author list * Address further review comments * Fix documentation * Last changes to administration/contribution * Use RTD theme
-
- Sep 26, 2022
-
-
Lund, Amanda authored
- Reset reentrant state in ORANGE - Force boundary crossing if we can't get off a surface in FieldPropagator
-
- Sep 25, 2022
-
-
Lund, Amanda authored
- Allow energy, position, and direction to be sampled from distributions - Allow multiple particle types - Add a method to create a primary generator using some predefined energy, spatial, and angular distributions that can be specified in the json input
-
- Sep 24, 2022
-
-
Johnson, Seth R. authored
-
- Sep 23, 2022
-
-
Johnson, Seth R. authored
* Add test of stuck track (Amanda Lund <alund@anl.gov>) * Add next-surface info * Return zero distance when finding reentrant next step * Add changedir-on-boundary field propagator test * Recover from reentrant substep * Fix missing boundary assignment
-
- Sep 19, 2022
-
-
Lund, Amanda authored
* Fix edge case in field propagator where track takes a small step from a boundary * Fix case where reported distance traveled in field propagator could be larger than step
-
- Sep 16, 2022
-
-
Lund, Amanda authored
* Index micro xs builders by element component ID instead of element ID * Sort imported element components by element ID * Don't import Klein-Nishina micro xs (not sure why Geant4 builds element selectors for this model, but we don't need it since no element is sampled for the discrete interaction) * Update ImportedProcesses test for master (micro xs builder checks will fail) * Update ImportedProcesses test for fix
-
- Sep 14, 2022
-
-
Johnson, Seth R. authored
* Add diagnostic stepper * Add failing gamma test in pathological field * Return "success" with NaN instead of iterating to convergence * Fix near-infinite loop by traveling slightly beyond chord end point This is probably the right thing to do because an intersection at that point is still "good enough" to call a boundary crossing.
-
- Sep 13, 2022
-
-
Johnson, Seth R. authored
* Separate "JSON reading" from "params reading" * Share reals among all orange objects
-
- Sep 07, 2022
-
-
Johnson, Seth R. authored
* Rename stepper result * Update cmake presets * Define rsqrt * Minor consistency changes to driver and test * Add MagFieldEquation test * Delete default initializers * More minor refactoring * Use detail::ax for updating results * Uniformly use ints for explicit 3-element looping
-
Stefano Tognini authored
* Add SB tables to ImportData * Fix RootInterfaceLinkDef; Add SB map to celer-dump-data * Import Livermore PE data to ROOT file * Add atomic relaxation data to ROOT file * Update celer-dump-data * Expand RootImporter test harness; expand celeritas-dump-data; update test root files * Reduce code duplication * Add assertion macros * Add all-element reader (Seth) * Remove zombie code * Address Amanda's comments; simplify ImportData operator bool * Fix ImportProcessConverter bug; Add ImportData test harness * Simplify RootImporter test * Update four-steel-slabs.root file: no reader data is stored * Update app/data/four-steel-slab.gdml to match the one in test/celeritas/data to avoid confusion; Update respective org.json accordingly * Refactor micro xs ElementPhysicsVectorMap -> ElementPhysicsVectors * Move content of ImportData.test into GeantImporter.test * Address comments
-
Johnson, Seth R. authored
* Add presets for summit * Reset describe when rebuilding from presets * Loosen test tolerances Compiling with -march=native -mtune=native on Power9 gives slightly different answers. VecGeom and ORANGE have slightly different answers, and using -fastmath gives *very* different answers (not yet accounted for in our test comparison so this still fails). * Fix geo-check test when run outside of source dir * Hide scary host code from timid NVCC * Fix missing operator bool on cu/hip RNG * Turn off fast math by default * Fix attempt to build unsupported muon processes when ROOT is disabled * Die correctly when attempting to build Coulomb process * Use celeritas basic physics list by default * Fix output of geant options * Enable geo check build if vecgeom is disabled * Fix duplicate main definitino * Add hidden variable for forcing resource lock on tests
-
- Sep 05, 2022
-
-
Johnson, Seth R. authored
* Add exception output * Add test of build output * Add more detailed output for exceptions
-
Johnson, Seth R. authored
* Add rewriting script * Apply script * Update everything and clang-format 14.0.6 * Delete tautological using statements * Fix cuda build * Move testing details to testdetail dir/namespace I think this makes it easier to find where all the components are as well as resolving @pcanal's and my concerns about ambiguity :) * Move main to TestMain * Remove unneeded celeritas:: namespaces in test * Move field details into general test namespace
-
- Sep 03, 2022
-
-
Johnson, Seth R. authored
-
- Aug 29, 2022
-
-
Johnson, Seth R. authored
* Move Geant detector construction to a different file * Add StringEnumMap * Allow more granular physics construction through demo loop * Move small classes into GeantSetup * Add JSON I/O for nonnative units * Add additional options to Geant setup, and update test * Fix unique_ptr for geant v10 * Fix VecGeom test * Fix energy limits * Refactor micro xs export * Add option to print geant4 setup options * Add command to regenerate ROOT file * Adjust root importer formatting * Add support for coulomb scattering import and clean up import process converter
-
- Aug 27, 2022
-
-
Johnson, Seth R. authored
* Add git-blame-ignore-revs file See https://docs.github.com/en/repositories/working-with-files/using-files/viewing-a-file#ignore-commits-in-the-blame-view and https://michaelheap.com/git-ignore-rev/ ``` git config --global blame.ignoreRevsFile .git-blame-ignore-revs git config --global blame.markIgnoredLines true git config --global blame.markUnblamableLines true ``` * Update documentation and mark TM as draft * Add developer and contributor documentation * Apply suggestions from code review * Apply suggestions from code review * Incorporate additional feedback and make a few more tweaks * Add release notes Co-authored-by: Philippe Canal <pcanal@fnal.gov> Co-authored-by: Amanda Lund <alund@anl.gov> Co-authored-by: Paul Romano <paul.k.romano@gmail.com>
-
- Aug 26, 2022
-
-
Johnson, Seth R. authored
* Fix three-spheres GDML file * REVERTME: debug code for failing heuristic track * Add failing orange test * Add surface normal accessor to simple unit tracker * Check for sign change in dot product with normal when changing direction on boundary * Revert "REVERTME: debug code for failing heuristic track" * Update geo heuristic path lengths with ORANGE * Update results and tolerances for vecgeom * Simplify logic via review feedback
-
- Aug 25, 2022
-
-
Canal, Philippe authored
* Make sure that that we use a consistent cuda runtime) * Remove 'Force shared cuda libraries when linking against vecgeom' * Extend comment on 'un-annotated' cuda depend libraries * Make sure that that we use a consistent cuda runtime also for executable * CMake/CUDA: refactor and fix the libceleritas_io case i.e the case of a library that depend of CUDA code but does not in itself include any cuda code.
-
Johnson, Seth R. authored
* Add geometry test * Add an assertion * Add geo.is_on_surface geometry function * Fix vecgeom errors by setting boundary state * Remove unnecessarily loose tolerance Reverts half of #493 * Don't warn about missing matids for 'unused' geometry volume labels * Add vecgeom version to diagnositc output * Correct typos, improve documentation and tests * Add new assertions based on 'is_on_surface' * Add another boundary assertion * Make sure next and max steps match when not hitting boundary * is_on_{surface -> boundary} * Convert some checks to is_on_boundary * Review feedback * Update test results for vecgeom, and disable ORANGE since it fails on reentrant tracks
-
- Aug 23, 2022
-
-
Johnson, Seth R. authored
* Work around floating point error in FieldPropagator * Add tests for step endpoint being near boundary * Reduce reported traveled distance when substep end point is close to boundary * Fix comments
-
Johnson, Seth R. authored
* Do not displace very small distances * Reformulate safety/displacement clamping * Don't displace for geo-limited steps * Fix accidental enabling of MSC when enable_msc=false in demo loop * Fix AlongStepUniformMscAction accessor
-
Lund, Amanda authored
* Make the imported model micro xs more consistent with Geant4 -Only import model micro xs for models that require it -Fix the energy grid lower bound -Expand test * Use secondary production cut in model micro xs calculation * Update test results * Recalculate range instead of hardcoding in UrbanMsc test * Check inverse Tsai radiation length in RootImporter test so differences between the very small rad length values will be caught with soft equal comparison * Call SetupForMaterial() before calculating micro xs in ImportProcessConverter and update root files and test again * Address review feedback * Revert me later: don't import e_coulomb_scattering micro xs until we can set secondary production cuts correctly for filtered secondary particles * Use G4KleinNishinaCompton instead of G4KleinNishinaModel; update root files again
-
- Aug 22, 2022
-
-
Johnson, Seth R. authored
-
Johnson, Seth R. authored
* Revert "Force disabling of shared libs for newer vecgeom (#489)" This reverts commit 37473126. * Include cuda headers instead of linking against cudart * Use an interface library to propagate toolkit includes * Remove stack size adjustment to fix occasional OOM * Force shared cuda libraries when linking against vecgeom * Update stepper tests for wildstyle
-
- Aug 21, 2022
-
-
Johnson, Seth R. authored
The unintentional explicit constructor causes problems with nljson on clang 14. This showed up because of the IO in #463.
-
- Aug 19, 2022
-
-
Guilherme Lima authored
* Loosen up real_type comparison in assert * Change extra_push() value to 1e-13 * Make the condition more general concerning float vs. double
-
Lund, Amanda authored
Set remaining model lower limits to zero and make sure demo loop runs with integral approach off (#490) * Set mu brems model lower energy limit to zero * Relax assertion to allow momentum_sq == 0 * Move BhabhaEnergyDistribution.hh to em/distribution * Remove second Moller-Bhabha model limit - Remove MB model min_valid_energy in MollerBhabhaData - Remove assertion that electron cutoff > model min_valid_energy in MB interactor
-
- Aug 18, 2022
-
-
Soon Yung Jun authored
* Fix the lowest kinetic energy to sample the multiple scattering * Return the original direction when the msc sampling is skipped
-
- Aug 17, 2022
-
-
Johnson, Seth R. authored
-
Lund, Amanda authored
* Return one vector per event from EventReader and PrimaryGenerator * Fix EventReader::operator() return type with HepMC3 disabled
-
Guilherme Lima authored
* Bring latest upstream updates to the BVHNavigator * Updating URL to reference source code * Remove decimal periods to avoid undesirable single to double promotions
-
- Aug 16, 2022
-
-
Johnson, Seth R. authored
* Backport std::unreachable * Fix unreachable warning * Clean additional warnings * Fix build and apply suggestion * Fix new warning from gcc * Apply feedback * Only add no-unknown-pragmas where supported
-
- Aug 12, 2022
-
-
Johnson, Seth R. authored
* Add process builder * Build ldemo processes with processbuilder
-
- Aug 11, 2022
-
-
Johnson, Seth R. authored
* Fix unintentional copy * Allow empty EM parameters * Fix process converter for particles not stored in cutoffs * Downgrade duplicate process/model warnings Duplicates are expected because of how we insert the data * Add filtering to GeantImporter and test This also reuses particle/element/material data rather than reloading it.
-
- Aug 09, 2022
-
-
Johnson, Seth R. authored
* Update field propagator test based on discussion * Test boundary crossing * Add nullop surface label printing to vecgeom * Add tests for surface crossings and corners
-
- Aug 08, 2022
-
-
Johnson, Seth R. authored
* Change order of operations to be more constexpr friendly * Use fastpow instead of pow * Be more consistent about units
-
Johnson, Seth R. authored
* Add tests for previously failing range-to-step * Replace output with actual test * Change 1e-6 to separate factors One is a multiplicative factor (unitless) and the other is additive (units of length).
-
Johnson, Seth R. authored
Don't evaluate the step calculation equation for small range $r$ (just above the cutoff $\rho$) but instead return $s = r$, since for $r = \rho + \epsilon$, the step equation simplifies to $s = \rho(1 + \epsilon) + O(\epsilon^2) \sim r$).
-
- Aug 03, 2022
-
-
Johnson, Seth R. authored
* Add some field propagator assertions * Gut field tests * Add more tests, two misbehaving * Add layered test too * Fix world definition * Re-add revolutions through CMS field * Fix max_nsteps return type * Revert clamp: now correct min(max(a, b), c) Fixes an error I introduced in #399 (but at least the assertion lets us know the precondition was violated)
-