Josh Bleecher Snyder
a53e38424c
base58: allocate less in Encode
...
* calculate maximum output length more precisely
to avoid allocation in the append
* use big.Int.Sign instead of needing bigZero
name old time/op new time/op delta
Base58Encode_5K-8 5.86ms ± 3% 5.79ms ± 2% -1.27% (p=0.035 n=9+10)
Base58Encode_100K-8 2.23s ± 1% 2.23s ± 0% ~ (p=0.074 n=9+8)
Base58Decode_5K-8 281µs ± 1% 282µs ± 1% ~ (p=0.720 n=9+10)
Base58Decode_100K-8 89.4ms ± 7% 88.3ms ± 7% ~ (p=0.123 n=10+10)
name old speed new speed delta
Base58Encode_5K-8 854kB/s ± 3% 864kB/s ± 2% ~ (p=0.134 n=9+10)
Base58Encode_100K-8 40.0kB/s ± 0% 40.0kB/s ± 0% ~ (all equal)
Base58Decode_5K-8 24.3MB/s ± 1% 24.2MB/s ± 1% ~ (p=0.644 n=9+10)
Base58Decode_100K-8 1.53MB/s ± 7% 1.55MB/s ± 7% ~ (p=0.218 n=10+10)
name old alloc/op new alloc/op delta
Base58Encode_5K-8 28.7kB ± 0% 19.2kB ± 0% -33.03% (p=0.000 n=10+10)
Base58Encode_100K-8 557kB ± 0% 385kB ± 0% -30.88% (p=0.000 n=10+10)
Base58Decode_5K-8 349kB ± 0% 349kB ± 0% ~ (all equal)
Base58Decode_100K-8 133MB ± 0% 133MB ± 0% ~ (p=0.183 n=10+10)
name old allocs/op new allocs/op delta
Base58Encode_5K-8 5.00 ± 0% 4.00 ± 0% -20.00% (p=0.000 n=10+10)
Base58Encode_100K-8 5.00 ± 0% 4.00 ± 0% -20.00% (p=0.000 n=10+10)
Base58Decode_5K-8 129 ± 0% 129 ± 0% ~ (all equal)
Base58Decode_100K-8 2.51k ± 0% 2.51k ± 0% ~ (p=0.321 n=10+10)
When Go 1.16 is released, performance will improve
significantly due to improvements to math/big.Int's division implementation.
2020-12-08 09:37:02 -05:00
Egon Elbre
594180da5d
base58: optimize Decode
...
Before:
BenchmarkBase58Decode_5K-32 266 4373774 ns/op 1.56 MB/s
BenchmarkBase58Decode_100K-32 1 1516196700 ns/op 0.09 MB/s
After:
BenchmarkBase58Decode_5K-32 3868 277944 ns/op 24.57 MB/s
BenchmarkBase58Decode_100K-32 13 83772100 ns/op 1.63 MB/s
2020-11-24 13:21:44 -05:00
Egon Elbre
ed1fc7ad99
base58: optimize Encode
...
Before:
BenchmarkBase58Encode_5K-32 46 23934763 ns/op 0.21 MB/s
BenchmarkBase58Encode_100K-32 1 9351948600 ns/op 0.01 MB/s
After:
BenchmarkBase58Encode_5K-32 501 2419129 ns/op 2.07 MB/s
BenchmarkBase58Encode_100K-32 2 923507950 ns/op 0.11 MB/s
2020-11-24 13:21:44 -05:00
David Hill
a5ecb5d954
travis: +go 1.8.x -go 1.6.x ( #94 )
2017-04-19 10:14:49 -04:00
Dave Collins
1b73e9828d
Relicense to the btcsuite developers.
2015-05-01 12:41:58 -05:00
Josh Rickmar
dca623d4ef
Optimize base58 decoding.
...
This change introduces an autogenerated base58 digit table to remove
the need to find the index of a character in the modified base58
alphabet each time. Additionally, it removes some unnecessary big
integer allocations to cut down on the GC churn.
Before:
BenchmarkBase58Encode 20 64998995 ns/op 0.08 MB/s
BenchmarkBase58Decode 50 35965928 ns/op 0.19 MB/s
Now:
BenchmarkBase58Encode 20 64644351 ns/op 0.08 MB/s
BenchmarkBase58Decode 200 7914748 ns/op 0.86 MB/s
2015-01-24 14:59:51 -05:00
benma
a72a54be16
Create a new sub-package base58.
...
It consists of the previous base58 implementation and new
base58 check encoding/decoding functions.
2014-12-22 11:43:07 +01:00
