Seth R. Johnson

Results 149 issues of Seth R. Johnson

*Derived from https://github.com/celeritas-project/docs/issues/17#issuecomment-2985727260 and associated dsicussion* Ideally we could launch an entire step without having to wait on a CUDA return result, so that we can do useful work on...

physics
performance

The parameter setup that's evolved in `accel` is fragile and half-duplicated in several other places in the code. We want migrate to using the `inp` data structures. UI integration: -...

app

Following on to #1974, no experiments or users have the custom Celeritas scintillation mean/stdev distributions: they all use grid-based inputs in energy space. So let's use those, but still convert...

enhancement
physics

The new vecgeom 2 solid navigator in #1720 is 3–5⨉ slower than v1.0 because it does a full relocation at each distance calculation, and those relocations are discarded most of...

performance
geometry

Currently the number of track slots has to be set manually, and increasing the number of Geant4 CPU threads (i.e., streams) offloading to Celeritas decreases the GPU performance because the...

core
performance

The primary push mechanism with the stepper takes a Span, resizes a collection, and copies (with a stream, but still usually synchronously since the span points to unpinned memory) primaries...

core
performance

We want users to be able to mark certain regions of the problem as requiring CPU (step sequential) processing. As a first step this can be a impl volume ->...

enhancement

In the future, due to ongoing work by @whokion , we expect a new Geant4 feature to allow G4 optical generation processes to emit distributions of photons rather than photons....

enhancement
external

We export many `CELERITAS_` variables to our installed `CeleritasConfig.cmake` file; many of them (e.g. CELERITAS_USE_X) found by introspecting the cache variable properties. However, in SCALE (and this is legal in...

bug

- Use a 2D grid for deduplication. Each dimension must be set up so that any two surfaces that are "soft equivalent" must be in the same *or neighboring* (including...

orange
performance