Skip to content

Add tagdepth fast path to tag comparison #807

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ChrisRackauckas wants to merge 1 commit into
base: master
Choose a base branch
from
+8 −2
Open

Add tagdepth fast path to tag comparison #807

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 8 additions & 2 deletions src/config.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,10 +20,16 @@ end

Tag(::Nothing, ::Type{V}) where {V} = nothing

@inline tagdepth(::Type) = 0
@inline tagdepth(::Type{<:Dual{T,V,N}}) where {T,V,N} = 1 + tagdepth(V)
@inline tagdepth(::Type{<:Tag{F,V}}) where {F,V} = 1 + tagdepth(V)
Comment on lines +23 to +25
Copy link
Copy Markdown
Member

@devmotion devmotion May 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why not just defining this function on ::Type and Type{<:Dual}? Why ::Type{<:Tag} as well?

Copy link
Copy Markdown
Member Author

@ChrisRackauckas ChrisRackauckas May 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Because when you nest it's a tag of duals


@inline function ≺(::Type{Tag{F1,V1}}, ::Type{Tag{F2,V2}}) where {F1,V1,F2,V2}
tagcount(Tag{F1,V1}) < tagcount(Tag{F2,V2})
end
d1 = tagdepth(Tag{F1,V1})
d2 = tagdepth(Tag{F2,V2})
d1 != d2 && return d1 < d2
Comment on lines +28 to +30
Copy link
Copy Markdown
Member

@devmotion devmotion May 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You can define Tags basically arbitrarily, so why would this be safe?

Copy link
Copy Markdown
Member Author

@ChrisRackauckas ChrisRackauckas May 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If the tags are defined according to the interfaces in the package, T is the type being differentiated. If T is the type being differentiated, then this gaurentees tag ordering by differentiation hierarchy. We previously had PRs closed about documenting this saying that it is non public API, so since this package always obeys that invarient internally and it's purposefully non public, it would be non breaking to enforce it.

return tagcount(Tag{F1,V1}) < tagcount(Tag{F2,V2})
Copy link
Copy Markdown
Member

@devmotion devmotion May 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It seems one could still run into the same precompilation-caused problems here, e.g., if V1 === V2 (e.g. both Float64)?

Copy link
Copy Markdown
Member Author

@ChrisRackauckas ChrisRackauckas May 13, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Build an example? All of the examples from before that cannot happen because it's not Float64 in both cases but Dual of Float64, and that tag nesting is exactly the part you missed from the earlier part.

end

struct InvalidTagException{E,O} <: Exception
end
Expand Down
Loading