dnsimple / base32_erlang / 14218427491

100%

push

github

Apply binary optimisations (#10)

* Apply binary optimisations

This is a port of the elixir implementation you can find at
https://hexdocs.pm/elixir/Base.html

Encoding is 8-12x faster, decoding 2-4x.

Some benchmarks using benchee:

Hardware:
Operating System: macOS
CPU Information: Apple M3 Pro
Number of Available Cores: 12
Available memory: 36 GB
Elixir 1.18.2
Erlang 27.3
JIT enabled: true

Encode:
```
Benchmark suite executing with the following configuration:
warmup: 2 s
time: 5 s
memory time: 5 s
reduction time: 5 s
parallel: 1
inputs: 17 bytes, 64 bytes
Estimated total run time: 1 min 8 s

Benchmarking candidate with input 17 bytes ...
Benchmarking candidate with input 64 bytes ...
Benchmarking original with input 17 bytes ...
Benchmarking original with input 64 bytes ...
Calculating statistics...
Formatting results...

----- With input 17 bytes -----
Name                ips        average  deviation         median         99th %
candidate       10.70 M       93.44 ns    ±62.56%          83 ns         125 ns
original         1.35 M      742.09 ns  ±2436.40%         625 ns        1000 ns

Comparison:
candidate       10.70 M
original         1.35 M - 7.94x slower +648.65 ns

Extended statistics:

Name              minimum        maximum    sample size                     mode
candidate            0 ns       96416 ns        19.06 M                    83 ns
original           333 ns    19270541 ns         5.67 M                   625 ns

Memory usage statistics:
Name         Memory usage
candidate           104 B
original            600 B - 5.77x memory usage +496 B

Reduction count statistics:
Name      Reduction count
candidate               8
original              138 - 17.25x reduction count +130

----- With input 64 bytes -----
Name                ips        average  deviation         median         99th %
candidate        6.35 M       0.157 μs    ±41.46%       0.166 μs        0.21 μs
original         0.56 M        1.78 μs   ±875.52%        1... (continued)

96 of 96 new or added lines in 1 file covered. (100.0%)

98 of 98 relevant lines covered (100.0%)

5493.31 hits per line