Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #813 +/- ##
==========================================
+ Coverage 70.53% 70.74% +0.21%
==========================================
Files 96 106 +10
Lines 5976 6710 +734
==========================================
+ Hits 4215 4747 +532
- Misses 1761 1963 +202
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Pull Request Overview
Aligns particle shifting and velocity extraction routines to be GPU-compatible by introducing specialized dispatch and buffer handling.
- Adds
current_velocityoverloads for GPU-friendly extraction based on density calculators. - Refactors
particle_shifting!to use a combined temp cache and replaces CPU-based max-velocity computation with a GPU-compatible approach. - Adjusts the literal for
Rto a rational form.
Reviewed Changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
| src/schemes/fluid/weakly_compressible_sph/system.jl | Added current_velocity methods for SummationDensity and ContinuityDensity |
| src/callbacks/particle_shifting.jl | Renamed temp cache to vu_cache, replaced eachparticle max with reinterpret approach, and changed R literal |
Comments suppressed due to low confidence (2)
src/schemes/fluid/weakly_compressible_sph/system.jl:254
- [nitpick] Consider adding unit tests for both
SummationDensityandContinuityDensitycases ofcurrent_velocityto verify GPU-compatible behavior under different density calculators.
@inline function current_velocity(v, ::SummationDensity,
src/callbacks/particle_shifting.jl:68
SVectoris used here but not imported in this file; addusing StaticArrays: SVectorat the top to prevent an undefined constant error.
reinterpret(reshape, SVector{ndims(system), eltype(v)},
|
\run-gpu-tests |
|
/run-gpu-tests |
As requested by @LasNikas.