Refactor scalar range getindex #50467
Merged
Conversation
This file contains 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
internal functional changes are - unsafe_getindex(::AbstractRange, ::Bool) no longer throws - removed @inline annotations
vtjnash
reviewed
Jul 11, 2023
github-merge-queue bot
pushed a commit
that referenced
this pull request
Jul 15, 2023
I noticed [here](#50467 (comment)) that `lastindex(x::Base.OneTo)` is not simply `x.stop`. This PR performs that optimization and many more by assuming `size` always returns positive numbers. ``` julia> @code_native lastindex(Base.OneTo(5)) # master .section __TEXT,__text,regular,pure_instructions .build_version macos, 13, 0 .globl _julia_lastindex_81 ; -- Begin function julia_lastindex_81 .p2align 2 _julia_lastindex_81: ; @julia_lastindex_81 ; ┌ @ abstractarray.jl:423 within `lastindex` ; %bb.0: ; %top ; │┌ @ abstractarray.jl:386 within `eachindex` ; ││┌ @ abstractarray.jl:134 within `axes1` ; │││┌ @ range.jl:708 within `axes` ; ││││┌ @ range.jl:471 within `oneto` ; │││││┌ @ range.jl:469 within `OneTo` @ range.jl:454 ; ││││││┌ @ promotion.jl:532 within `max` ; │││││││┌ @ int.jl:83 within `<` ldr x8, [x0] ; │││││││└ ; │││││││┌ @ essentials.jl:642 within `ifelse` cmp x8, #0 csel x0, x8, xzr, gt ; │└└└└└└└ ret ; └ ; -- End function .subsections_via_symbols julia> @code_native lastindex(Base.OneTo(5)) # pr .section __TEXT,__text,regular,pure_instructions .build_version macos, 13, 0 .globl _julia_lastindex_13253 ; -- Begin function julia_lastindex_13253 .p2align 2 _julia_lastindex_13253: ; @julia_lastindex_13253 ; ┌ @ abstractarray.jl:423 within `lastindex` ; %bb.0: ; %top ldr x0, [x0] ret ; └ ; -- End function .subsections_via_symbols ``` Also removed `axes(r::AbstractRange) = (oneto(length(r)),)` (added in #40382, @vtjnash) as redundant with the general `axes` method. The obvious downside here is that if someone defines an object with negative size, its axes will include Base.OneTo with negative stop. I think that is acceptable, but if not, we can gate this optimization to a set of known types (all AbstractArray types defined in Base should have non-negative size)
LilithHafner
commented
Jul 15, 2023
barucden
reviewed
Jul 17, 2023
Co-authored-by: Denis Barucic <[email protected]>
LilithHafner
commented
Jul 17, 2023
|
Oops, I forgot about this. |
LilithHafner
added
performance
|
@vtjnash found a performance regression due to this PR |
LilithHafner
added a commit
that referenced
this pull request
Sep 5, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
The point of this is to make these internals simpler and easier to edit going forward. I originally wanted to include these changes with #50118, but held back to expedite that bugfix.
Functional changes are
unsafe_getindex(::AbstractRange, ::Bool)no longer throws@inlineannotations, but the functions annotated with@inlineshould typically be below the inlining threshold anyway.