Fix: up to 80% speed and memory optimizations for slices

[w1kend]

Hello! 👋
I've made some optimizations for slices. There are also benchmarks to check the results.

And my results under:

go version go1.19 darwin/arm64
macbook air m1
goos: darwin
goarch: arm64

Flatten

name                    old time/op    new time/op    delta
Flatten/ints_10-8          294ns ± 1%     116ns ± 2%  -60.65%  (p=0.000 n=9+10)
Flatten/ints_100-8        23.7µs ± 0%     5.8µs ± 0%  -75.58%  (p=0.000 n=10+8)
Flatten/ints_1000-8       2.47ms ± 2%    0.63ms ± 1%  -74.58%  (p=0.000 n=10+8)
Flatten/strings_10-8       705ns ± 1%     227ns ± 0%  -67.83%  (p=0.000 n=10+9)
Flatten/strings_100-8     77.5µs ±15%    25.2µs ± 5%  -67.42%  (p=0.000 n=10+10)
Flatten/strings_1000-8    24.2ms ± 1%     5.5ms ± 2%  -77.46%  (p=0.000 n=10+10)

name                    old alloc/op   new alloc/op   delta
Flatten/ints_10-8         2.48kB ± 0%    0.90kB ± 0%  -63.87%  (p=0.000 n=10+10)
Flatten/ints_100-8         356kB ± 0%      82kB ± 0%  -77.00%  (p=0.000 n=10+10)
Flatten/ints_1000-8       44.9MB ± 0%     8.0MB ± 0%  -82.17%  (p=0.000 n=10+8)
Flatten/strings_10-8      5.09kB ± 0%    1.79kB ± 0%  -64.78%  (p=0.000 n=10+10)
Flatten/strings_100-8      683kB ± 0%     164kB ± 0%  -76.01%  (p=0.000 n=10+10)
Flatten/strings_1000-8    81.6MB ± 0%    16.0MB ± 0%  -80.39%  (p=0.000 n=9+10)

Chunk

name                  old time/op    new time/op    delta
Chunk/strings_10-8       120ns ± 0%      29ns ± 1%  -75.75%  (p=0.000 n=8+10)
Chunk/strings_100-8      880ns ± 0%     100ns ± 2%  -88.67%  (p=0.000 n=8+10)
Chunk/strings_1000-8    8.27µs ± 0%    0.68µs ± 1%  -91.81%  (p=0.000 n=9+9)
Chunk/ints10-8          88.5ns ± 0%    29.1ns ± 0%  -67.09%  (p=0.000 n=9+8)
Chunk/ints100-8          546ns ± 0%      99ns ± 0%  -81.96%  (p=0.000 n=8+9)
Chunk/ints1000-8        4.97µs ± 0%    0.67µs ± 1%  -86.43%  (p=0.000 n=10+10)

Replace

name                      old time/op    new time/op    delta
Replace/strings_1000-8      5.24µs ± 0%    1.29µs ± 0%  -75.46%  (p=0.008 n=5+5)
Replace/strings_10000-8     53.2µs ± 1%    13.1µs ± 1%  -75.37%  (p=0.008 n=5+5)
Replace/strings_100000-8     822µs ± 1%     127µs ± 2%  -84.50%  (p=0.016 n=5+4)
Replace/ints1000-8          1.50µs ± 0%    0.72µs ± 1%  -52.24%  (p=0.008 n=5+5)
Replace/ints10000-8         12.6µs ± 0%     3.8µs ± 5%  -69.93%  (p=0.016 n=4+5)
Replace/ints100000-8         126µs ± 1%      30µs ± 1%  -76.44%  (p=0.008 n=5+5)

Drop funcs - 10-30%

[samber]

Oops, thanks for these obvious optimization!