Add Seltzer-Berger interactor and kernel #241
Conversation
|
Looking for first round of reviews while documentation and test updated, and add Modified Tsai sampler for bremsstrahlung and pair-production. |
| cutoffs_.energy(device_pointers_.ids.gamma), | ||
| inc_energy_); | ||
| gamma_exit_energy | ||
| = Energy{gamma_exit_energy.value() * scale_xs(gamma_exit_energy)}; |
There was a problem hiding this comment.
This block isn't quite right (although we'll need to use the constructor you've enumerated). The XsCorrector needs to be used as part of the energy distribution, which I didn't implement because I didn't want to interfere with your ongoing work on the SB interactor 😬 . The XsCorrector should modify the energy distribution to be sampled from, rather than scaling the exiting energy as written here.
vrpascuzzi
force-pushed
the
feature/physics/SeltzerBergerInteractor
branch
from
78dd506
to
b2f848d
Compare
|
Hope to have addressed most of the reviews. Will continue with remaining documentation and tests Monday. |
This commit adds the SB interactor and associated kernel method.
* Initialize `material_view` near where first used
* `migdal` as `constexpr`, and use integer multiplication to
prevent float->double conversion
* Make type aliases `public` and add doxygen mark-up * Remove unused `energy_val_max_`
* Add ElementId class member to interactor and supply during
instantiation of interactor in CUDA source
* Remove assertion on minimum energy for interactor
* Do not apply XS correction for positrons yet in interactor;
to be integrated in energy distribution sampler
* Remove `energy_val_min_` member from interactor (handled
before interactor is constructed); instead, check incoming
particle energy is sufficient to produce secondary at start
of `operator()`
Used for sampling angular distributions in pair-production and bremsstrahlung processes (e.g. Seltzer-Berger and Bethe-Heitler models).
vrpascuzzi
force-pushed
the
feature/physics/SeltzerBergerInteractor
branch
from
b2f848d
to
71a3b94
Compare
* Update class description and other comments
* Remove photon angular sampler function (using instead common
TsaiUrbanDistribution)
Sampled electron angle twice. Fixed to sample positron.
vrpascuzzi
force-pushed
the
feature/physics/SeltzerBergerInteractor
branch
from
1e8fd0f
to
3e8470f
Compare
* Use native pointers/refs \todo Fix issue with min_gamma_energy <= 0
| @@ -149,34 +150,20 @@ CELER_FUNCTION Interaction BetheHeitlerInteractor::operator()(Engine& rng) | |||
| real_type phi = sample_phi(rng); | |||
|
|
|||
| // Electron | |||
| real_type cost = this->sample_cos_theta(secondaries[0].energy.value(), rng); | |||
| TsaiUrbanDistribution sample_electron_angle( | |||
| secondaries[0].energy, units::MevMass{shared_.inv_electron_mass}); | |||
There was a problem hiding this comment.
The units here are wrong but it looks like the quantity is right. The sampler and its G4ModifiedTsai counterpart both expect 1/(m_e c^2) so units of MevMassInv, which you can define (for now, and I'd dprob put it in the TsaiUrbanDistribution) as
using MevMassInv = Quantity<UnitDivide<CLightSq, Mev>>;* Explicitly mark public of type aliases and doxygen
* EXPECT_VEC_SOFT_EQ for angle comparison with `double` type
* Store ratio energy [MeV] to mass [MevMass*c^2] instead
of energy and mass separately
|
SB test still failing due to EDIT: Fixed, thanks to @sethrj. Finish test and wrap up tomorrow! |
* `gamma_secondary` -> `secondaries` (for consistency) * Assign secondary energy * normalize incident particle outgoing direction
* Finalize 'basic' test * Add 'stress' test
A bit late to the party but we have it now! |
|
I'm gonna implement the positron XS correction; and then you can add the new Brems process. We shouldn't end up stepping on each other's changes. |
This commit adds the SB interactor and associated kernel method.