Repository navigation
Preserve floating reference dtype in transport matrix builders - #213
Merged
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
A float32 reference in
build_advection_matrixorbuild_diffusion_matrixpreviously produced float64 operators on NumPy and JAX with x64 enabled. Floating references now select their own precision directly; integer references still construct floating operators. The existing stencil formulas and namespace selection are preserved.Independent reflecting stencils cover NumPy/JAX/CPU Torch in float32/float64. JAX JIT and JAX/Torch coefficient derivatives match a known quadratic objective; the existing CPU Torch CI job includes these regressions. Documentation and changelog describe reference-controlled precision.
Closes #212. Found during #206 qualification.
Validation:
just format,just ci(742 core tests passed; 389 provider tests passed),just docs, the all-backend precision suite (22 passed), andjust build-allpassed, including both installed wheels. Required GPU qualification remains separate; this fix preserves the existing floating-reference API contract.