fix: promote the output array of initialization to the tunable eltype#4469
Merged
Conversation
SebastianM-C
force-pushed
the
smc/init_promotion
branch
from
2053909 to
f83c8e4
Compare
AayushSabharwal
approved these changes
Apr 27, 2026
Member
AayushSabharwal
left a comment
AayushSabharwal
left a comment
There was a problem hiding this comment.
I merged #4459 and this needs a rebase now 😅 but otherwise looks good.
SebastianM-C
force-pushed
the
smc/init_promotion
branch
2 times, most recently
from
4faee0d to
ddc82e4
Compare
Member
Author
|
rebased |
Member
|
This PR introduces failures in the initialization test group |
SebastianM-C
force-pushed
the
smc/init_promotion
branch
from
ddc82e4 to
33acf53
Compare
This adresses the cases where the we generate output arrays from fully constant RHS, which would end up being incompatible with the elype of parameters in ForwardDiff and other similar contexts. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-authored-by: Aayush Sabharwal <aayush.sabharwal@gmail.com>
Co-authored-by: Aayush Sabharwal <aayush.sabharwal@gmail.com>
SebastianM-C
force-pushed
the
smc/init_promotion
branch
from
33acf53 to
65b1ac4
Compare
Hardcoding Float64 silently promoted Float32 problems to Float64. Threading floatT preserves user-chosen precision while still rescuing integer RHS, and the dynamic tunable eltype read from nlsol still wins for ForwardDiff. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This adresses the cases where the we generate output arrays from fully constant RHS, which would end up being incompatible with the elype of parameters in ForwardDiff and other similar contexts.
Checklist
contributor guidelines, in particular the SciML Style Guide and
COLPRAC.
Additional context
After digging through what's going on in #4457 with Claude I think that I found a fix. The issue is that we generate observed functions that hardcode
Vector{Float64}regardless of the eltype of the tunables which ends up defeating the u0 promotion mechanisms that we have.This should fix #4457 and probably #3924 too, I think that the underlying issue is the same, but #4457 is the simpler reproducer.