Regression in broadcast assignment to a SlowSubArray on nightly
#53158
Labels
domain:arrays
[a, r, r, a, y, s]
kind:regression
Regression in behavior compared to a previous version
performance
Must go faster
regression 1.11
Comments
jishnub
added
performance
|
Bisected to 9aa7980: On this commit, julia> a = zeros(4000,4000); b = rand(size(a)...);
julia> @btime $a[1:end,1:end] .= $b;
61.351 ms (0 allocations: 0 bytes)vs on 045b6f9: julia> @btime $a[1:end,1:end] .= $b;
20.189 ms (0 allocations: 0 bytes) |
vtjnash
pushed a commit
that referenced
this issue
Feb 20, 2024
…3383) With this, the following (and equivalent calls) work: ```julia julia> copyto!(view(zeros(BigInt, 2), 1:2), Vector{BigInt}(undef,2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef julia> copyto!(view(zeros(BigInt, 2), 1:2), view(Vector{BigInt}(undef,2), 1:2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef ``` Close #53098. With this, all the `_unsetindex!` branches in `copyto_unaliased!` work for `Array`-views, and this makes certain indexing operations vectorize and speed-up: ```julia julia> using BenchmarkTools julia> a = view(rand(100,100), 1:100, 1:100); b = view(similar(a), axes(a)...); julia> @Btime copyto!($b, $a); 16.427 μs (0 allocations: 0 bytes) # master 2.308 μs (0 allocations: 0 bytes) # PR ``` Improves (but doesn't resolve) #40962 and #53158 ```julia julia> a = rand(40,40); b = rand(40,40); julia> @Btime $a[1:end,1:end] .= $b; 5.383 μs (0 allocations: 0 bytes) # v"1.12.0-DEV.16" 3.194 μs (0 allocations: 0 bytes) # PR ``` ƒ Co-authored-by: Jameson Nash <vtjnash@gmail.com>
tecosaur
pushed a commit
to tecosaur/julia
that referenced
this issue
Mar 4, 2024
…liaLang#53383) With this, the following (and equivalent calls) work: ```julia julia> copyto!(view(zeros(BigInt, 2), 1:2), Vector{BigInt}(undef,2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef julia> copyto!(view(zeros(BigInt, 2), 1:2), view(Vector{BigInt}(undef,2), 1:2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef ``` Close JuliaLang#53098. With this, all the `_unsetindex!` branches in `copyto_unaliased!` work for `Array`-views, and this makes certain indexing operations vectorize and speed-up: ```julia julia> using BenchmarkTools julia> a = view(rand(100,100), 1:100, 1:100); b = view(similar(a), axes(a)...); julia> @Btime copyto!($b, $a); 16.427 μs (0 allocations: 0 bytes) # master 2.308 μs (0 allocations: 0 bytes) # PR ``` Improves (but doesn't resolve) JuliaLang#40962 and JuliaLang#53158 ```julia julia> a = rand(40,40); b = rand(40,40); julia> @Btime $a[1:end,1:end] .= $b; 5.383 μs (0 allocations: 0 bytes) # v"1.12.0-DEV.16" 3.194 μs (0 allocations: 0 bytes) # PR ``` ƒ Co-authored-by: Jameson Nash <vtjnash@gmail.com>
mkitti
pushed a commit
to mkitti/julia
that referenced
this issue
Mar 7, 2024
…liaLang#53383) With this, the following (and equivalent calls) work: ```julia julia> copyto!(view(zeros(BigInt, 2), 1:2), Vector{BigInt}(undef,2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef julia> copyto!(view(zeros(BigInt, 2), 1:2), view(Vector{BigInt}(undef,2), 1:2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef ``` Close JuliaLang#53098. With this, all the `_unsetindex!` branches in `copyto_unaliased!` work for `Array`-views, and this makes certain indexing operations vectorize and speed-up: ```julia julia> using BenchmarkTools julia> a = view(rand(100,100), 1:100, 1:100); b = view(similar(a), axes(a)...); julia> @Btime copyto!($b, $a); 16.427 μs (0 allocations: 0 bytes) # master 2.308 μs (0 allocations: 0 bytes) # PR ``` Improves (but doesn't resolve) JuliaLang#40962 and JuliaLang#53158 ```julia julia> a = rand(40,40); b = rand(40,40); julia> @Btime $a[1:end,1:end] .= $b; 5.383 μs (0 allocations: 0 bytes) # v"1.12.0-DEV.16" 3.194 μs (0 allocations: 0 bytes) # PR ``` ƒ Co-authored-by: Jameson Nash <vtjnash@gmail.com>
KristofferC
pushed a commit
that referenced
this issue
Mar 27, 2024
…3383) With this, the following (and equivalent calls) work: ```julia julia> copyto!(view(zeros(BigInt, 2), 1:2), Vector{BigInt}(undef,2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef julia> copyto!(view(zeros(BigInt, 2), 1:2), view(Vector{BigInt}(undef,2), 1:2)) 2-element view(::Vector{BigInt}, 1:2) with eltype BigInt: #undef #undef ``` Close #53098. With this, all the `_unsetindex!` branches in `copyto_unaliased!` work for `Array`-views, and this makes certain indexing operations vectorize and speed-up: ```julia julia> using BenchmarkTools julia> a = view(rand(100,100), 1:100, 1:100); b = view(similar(a), axes(a)...); julia> @Btime copyto!($b, $a); 16.427 μs (0 allocations: 0 bytes) # master 2.308 μs (0 allocations: 0 bytes) # PR ``` Improves (but doesn't resolve) #40962 and #53158 ```julia julia> a = rand(40,40); b = rand(40,40); julia> @Btime $a[1:end,1:end] .= $b; 5.383 μs (0 allocations: 0 bytes) # v"1.12.0-DEV.16" 3.194 μs (0 allocations: 0 bytes) # PR ``` ƒ Co-authored-by: Jameson Nash <vtjnash@gmail.com> (cherry picked from commit 1a90409)
|
This seems to have regressed on the current nightly ( julia> a = zeros(40000,4000); b = rand(size(a)...);
julia> @benchmark $a[1:end, 1:end] .= $b
BenchmarkTools.Trial: 16 samples with 1 evaluation.
Range (min … max): 311.599 ms … 332.538 ms ┊ GC (min … max): 0.00% … 0.00%
Time (median): 313.798 ms ┊ GC (median): 0.00%
Time (mean ± σ): 315.770 ms ± 5.354 ms ┊ GC (mean ± σ): 0.00% ± 0.00%
▁▁█▁▁▁██ ▁ ▁ ▁ ▁ ▁
████████▁▁▁█▁▁▁▁▁▁▁▁▁█▁█▁▁▁▁▁█▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█ ▁
312 ms Histogram: frequency by time 333 ms <
Memory estimate: 0 bytes, allocs estimate: 0.vs on nightly: julia> @benchmark $a[1:end, 1:end] .= $b
BenchmarkTools.Trial: 12 samples with 1 evaluation.
Range (min … max): 448.373 ms … 452.969 ms ┊ GC (min … max): 0.00% … 0.00%
Time (median): 450.255 ms ┊ GC (median): 0.00%
Time (mean ± σ): 450.305 ms ± 1.671 ms ┊ GC (mean ± σ): 0.00% ± 0.00%
█ ██ ██ █ █ █ █ █ █ █
█▁██▁██▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁█▁▁█▁▁▁▁▁▁█▁▁▁▁█▁▁▁▁▁▁▁▁▁▁▁█▁▁▁█▁▁▁▁▁█ ▁
448 ms Histogram: frequency by time 453 ms <
Memory estimate: 0 bytes, allocs estimate: 0.
julia> VERSION
v"1.12.0-DEV.528" |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
On v1.10.0
vs on v"1.11.0-DEV.1442" as well as the current master (d54a455)
versioninfo:
Curiously, profiling points to integer comparison checks while iterating over
CartesianIndicesto be the most expensive step:The text was updated successfully, but these errors were encountered: