Added IndexRange to remove checks in hot loops

[jorgecarleitao]

This improves performance of sort by removing checks from inside hot loops.

Backward incompatible changes

The trait Index is now available at arrow2::types::Index, instead of arrow2::array::Index

[codecov]

Codecov Report

Merging #247 (5fbe923) into main (3d123b3) will increase coverage by 0.12%.
The diff coverage is 73.68%.

@@            Coverage Diff             @@
##             main     #247      +/-   ##
==========================================
+ Coverage   76.69%   76.81%   +0.12%     
==========================================
  Files         229      230       +1     
  Lines       19687    19666      -21     
==========================================
+ Hits        15099    15107       +8     
+ Misses       4588     4559      -29     

Impacted Files	Coverage Δ
src/array/mod.rs	52.25% <ø> (ø)
src/array/specification.rs	70.96% <ø> (+28.99%)	⬆️
src/buffer/mutable.rs	91.76% <ø> (+0.24%)	⬆️
src/compute/sort/boolean.rs	89.47% <ø> (ø)
src/compute/sort/lex_sort.rs	82.14% <ø> (ø)
src/compute/sort/mod.rs	53.55% <ø> (ø)
src/compute/sort/primitive/indices.rs	100.00% <ø> (ø)
src/compute/sort/utf8.rs	100.00% <ø> (ø)
src/compute/take/mod.rs	88.48% <ø> (ø)
src/types/mod.rs	52.00% <ø> (ø)
... and 4 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3d123b3...5fbe923. Read the comment docs.

[sundy-li]

LGTM, do we have benchmark results vs native ranges?

[jorgecarleitao]

Thanks for the review. This has no impact in performance. I think it makes it easier to read and rules out the from_usize bit. Thus, merging :)

[jorgecarleitao]

sort-limit 2^10 f32     time:   [12.611 us 12.639 us 12.666 us]                                 
                        change: [-0.3425% +0.2052% +0.9419%] (p = 0.61 > 0.05)
                        No change in performance detected.
Found 14 outliers among 100 measurements (14.00%)
  6 (6.00%) low severe
  4 (4.00%) low mild
  1 (1.00%) high mild
  3 (3.00%) high severe

sort-limit 2^12 f32     time:   [25.693 us 25.705 us 25.717 us]                                 
                        change: [-0.6963% -0.2300% +0.0666%] (p = 0.31 > 0.05)
                        No change in performance detected.
Found 13 outliers among 100 measurements (13.00%)
  4 (4.00%) low severe
  4 (4.00%) low mild
  4 (4.00%) high mild
  1 (1.00%) high severe

sort-limit 2^14 f32     time:   [78.742 us 78.811 us 78.871 us]                                
                        change: [-0.8476% -0.5057% -0.1062%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 16 outliers among 100 measurements (16.00%)
  4 (4.00%) low severe
  6 (6.00%) low mild
  4 (4.00%) high mild
  2 (2.00%) high severe

sort-limit 2^16 f32     time:   [448.16 us 449.77 us 452.01 us]                                
                        change: [-0.8242% -0.4075% +0.0675%] (p = 0.07 > 0.05)
                        No change in performance detected.
Found 10 outliers among 100 measurements (10.00%)
  2 (2.00%) low severe
  3 (3.00%) low mild
  1 (1.00%) high mild
  4 (4.00%) high severe

Benchmarking sort-limit 2^18 f32: Warming up for 3.0000 s
Warning: Unable to complete 100 samples in 5.0s. You may wish to increase target time to 8.7s, enable flat sampling, or reduce sample count to 50.
sort-limit 2^18 f32     time:   [1.7144 ms 1.7167 ms 1.7198 ms]                                 
                        change: [-0.6946% -0.4104% -0.1233%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 18 outliers among 100 measurements (18.00%)
  4 (4.00%) low severe
  7 (7.00%) low mild
  3 (3.00%) high mild
  4 (4.00%) high severe

sort-limit 2^20 f32     time:   [9.0083 ms 9.0169 ms 9.0250 ms]                                
                        change: [+0.1708% +0.3120% +0.4439%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 11 outliers among 100 measurements (11.00%)
  5 (5.00%) low severe
  2 (2.00%) low mild
  2 (2.00%) high mild
  2 (2.00%) high sever