Skip to content

Conversation

oscardssmith
Copy link
Member

By my measurements, it's a bunch faster for this workload (this may be new as of Julia 1.11)

julia> function f3(d, n)
           for i in 1:n
               push!(d, i)
               pop!(d)
           end
       end
f3 (generic function with 1 method)

julia> @benchmark f3(d, 1000) evals=1 setup = (d=Vector{Int}())
BenchmarkTools.Trial: 10000 samples with 1 evaluation per sample.
 Range (min … max):  789.000 ns …  10.042 μs  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     797.000 ns               ┊ GC (median):    0.00%
 Time  (mean ± σ):   823.878 ns ± 158.991 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▅█▅▁▅▄           ▂  ▁▄▃                                     ▂ ▁
  ███████▆▆▄▁▁▁▁▁▁▇██▃███▇▇▆▇█▆▆▅▄▄▄▄▄▅▄▅▄▆▇▇▆▆▇▆▆▆▅▅▄▁▄▁▃▃▄▅██ █
  789 ns        Histogram: log(frequency) by time       1.07 μs <

 Memory estimate: 96 bytes, allocs estimate: 1.

julia> @benchmark f3(d, 1000) evals=1 setup = (d=Deque{Int}())
BenchmarkTools.Trial: 10000 samples with 1 evaluation per sample.
 Range (min … max):  3.667 μs …  28.349 μs  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     3.803 μs               ┊ GC (median):    0.00%
 Time  (mean ± σ):   3.930 μs ± 489.972 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▁▆█▄▇▅▃▃▂▁▁▂▅▃ ▁  ▁▁       ▁                                ▂
  ████████████████▇▇██▇▇▆▇█▇███▇▇▆█▇▄▅▄▆▆▆▆▇▇▅▅▅▄▅▅▆▅▄▅▄▄▄▄▆▅ █
  3.67 μs      Histogram: log(frequency) by time      5.66 μs <

By my measurements, it's a bunch faster for this workload (this may be new as of Julia 1.11)
```
julia> function f3(d, n)
           for i in 1:n
               push!(d, i)
               pop!(d)
           end
       end
f3 (generic function with 1 method)

julia> @benchmark f3(d, 1000) evals=1 setup = (d=Vector{Int}())
BenchmarkTools.Trial: 10000 samples with 1 evaluation per sample.
 Range (min … max):  789.000 ns …  10.042 μs  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     797.000 ns               ┊ GC (median):    0.00%
 Time  (mean ± σ):   823.878 ns ± 158.991 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▅█▅▁▅▄           ▂  ▁▄▃                                     ▂ ▁
  ███████▆▆▄▁▁▁▁▁▁▇██▃███▇▇▆▇█▆▆▅▄▄▄▄▄▅▄▅▄▆▇▇▆▆▇▆▆▆▅▅▄▁▄▁▃▃▄▅██ █
  789 ns        Histogram: log(frequency) by time       1.07 μs <

 Memory estimate: 96 bytes, allocs estimate: 1.

julia> @benchmark f3(d, 1000) evals=1 setup = (d=Deque{Int}())
BenchmarkTools.Trial: 10000 samples with 1 evaluation per sample.
 Range (min … max):  3.667 μs …  28.349 μs  ┊ GC (min … max): 0.00% … 0.00%
 Time  (median):     3.803 μs               ┊ GC (median):    0.00%
 Time  (mean ± σ):   3.930 μs ± 489.972 ns  ┊ GC (mean ± σ):  0.00% ± 0.00%

  ▁▆█▄▇▅▃▃▂▁▁▂▅▃ ▁  ▁▁       ▁                                ▂
  ████████████████▇▇██▇▇▆▇█▇███▇▇▆█▇▄▅▄▆▆▆▆▇▇▅▅▅▄▅▅▆▅▄▅▄▄▄▄▆▅ █
  3.67 μs      Histogram: log(frequency) by time      5.66 μs <

```
asinghvi17 and others added 4 commits April 30, 2025 21:47
this will just work with a vector I think
Co-authored-by: Anshul Singhvi <[email protected]>
Co-authored-by: Anshul Singhvi <[email protected]>
Remove outdated Iterators.reverse definition
@fingolfin
Copy link
Member

tests fail

Co-authored-by: Max Horn <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants