from_slice behavior

[horasal]

from_slice behavior

[horasal]

Thanks for your feedback!

I agree with you that current from_slice imprementation is inconsistent. Slices with repeated indices may cause it use more memory than expected.

We have several options to solve this:

1. Trace the max index and add a shirk_to before return.
2. Change initial length to slice.max() to avoid any allocation.
3. Change initial length to 0 make sure no memory is wasted.

I did some benches on these solutions, and it shows that the first one has better performance over others.
I'll make a patch later with the first solution and thanks for your feedback again!

from_slice, small set, capacity=slice.len()                                                                             
                        time:   [6.7193 us 6.7212 us 6.7232 us]
                        change: [+0.6780% +0.9091% +1.1639%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 21 outliers among 100 measurements (21.00%)
  3 (3.00%) low severe
  5 (5.00%) low mild
  5 (5.00%) high mild
  8 (8.00%) high severe

from_slice, small set, capacity=slice.max()                                                                             
                        time:   [6.8733 us 6.8753 us 6.8776 us]
                        change: [-0.2958% -0.1312% +0.0648%] (p = 0.18 > 0.05)
                        No change in performance detected.
Found 11 outliers among 100 measurements (11.00%)
  1 (1.00%) low mild
  3 (3.00%) high mild
  7 (7.00%) high severe

from_slice, small set, capacity=0                                                                             
                        time:   [6.8792 us 6.8811 us 6.8833 us]
                        change: [-0.3338% -0.0785% +0.1143%] (p = 0.56 > 0.05)
                        No change in performance detected.
Found 6 outliers among 100 measurements (6.00%)
  2 (2.00%) high mild
  4 (4.00%) high severe

from_slice, large set, capacity=slice.len()                                                                            
                        time:   [121.72 us 121.84 us 121.95 us]
                        change: [+0.3143% +0.4387% +0.5645%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high severe

from_slice, large set, capacity=slice.max()                                                                            
                        time:   [123.99 us 124.12 us 124.24 us]
                        change: [-0.4949% -0.3700% -0.2379%] (p = 0.00 < 0.05)
                        Change within noise threshold.

from_slice, large set, capacity=0                                                                            
                        time:   [124.02 us 124.15 us 124.28 us]
                        change: [-0.4067% -0.2901% -0.1640%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 1 outliers among 100 measurements (1.00%)
  1 (1.00%) high severe