Much faster/cheaper b58 ops

[ribasushi]

Adds an extra benchmark and dependency on external work by @mr-tron and myself, instead of copying code. Supersedes/closes https://github.com/btcsuite/btcutil/pull/93

Provides significant time/space reduction in address encoding/decoding

• Before

  $ go test ./base58/ -bench=. -benchmem
  goos: darwin
  goarch: amd64
  pkg: github.com/btcsuite/btcutil/base58
  BenchmarkBase58EncodeAddr-8       195880              5864 ns/op           5.80 MB/s         648 B/op         51 allocs/op
  BenchmarkBase58DecodeAddr-8       290460              4115 ns/op          11.42 MB/s         264 B/op          8 allocs/op
  BenchmarkBase58Encode5k-8             18          63641483 ns/op           0.08 MB/s       82440 B/op       6725 allocs/op
  BenchmarkBase58Decode5k-8            226           5334819 ns/op           1.28 MB/s     1024873 B/op        380 allocs/op

  • After

    $ go test ./base58/ -bench=. -benchmem
    goos: darwin
    goarch: amd64
    pkg: github.com/btcsuite/btcutil/base58
    BenchmarkBase58EncodeAddr-8       857293              1362 ns/op          24.97 MB/s          96 B/op          2 allocs/op
    BenchmarkBase58DecodeAddr-8      2099786               566 ns/op          83.08 MB/s         128 B/op          2 allocs/op
    BenchmarkBase58Encode5k-8             45          26290144 ns/op           0.19 MB/s       13824 B/op          2 allocs/op
    BenchmarkBase58Decode5k-8            147           8173594 ns/op           0.84 MB/s       17152 B/op          2 allocs/op

[ribasushi]

I apologize, my IDE screwed me over. New push incoming.

[ribasushi]

@Roasbeef @guggero any word on that PR? Anything I need to adjust here or in the imported library? Thanks!

[ribasushi]

Actually I had another insight, which brings the "after" to

pkg: github.com/btcsuite/btcutil/base58
BenchmarkBase58EncodeAddr-8      2690395           436 ns/op      77.94 MB/s          96 B/op          2 allocs/op
BenchmarkBase58DecodeAddr-8      3717728           322 ns/op     146.10 MB/s         128 B/op          2 allocs/op
BenchmarkBase58Encode5k-8            224       5261993 ns/op       0.95 MB/s       13824 B/op          2 allocs/op
BenchmarkBase58Decode5k-8            512       2300409 ns/op       2.97 MB/s       15104 B/op          2 allocs/op

I need to stabilize the new code a bit, and upstream it. Will open a new PR when ready.