Avoid function call overhead This is not a good solution, but it'll do for the moment. noasm: benchmark old MB/s new MB/s speedup BenchmarkHashes/xxhash,n=5_B-4 383.10 596.40 1.56x BenchmarkHashes/xxhash,n=100_B-4 3558.30 4301.40 1.21x BenchmarkHashes/xxhash,n=4_KB-4 8778.69 8840.83 1.01x BenchmarkHashes/xxhash,n=10_MB-4 7731.16 7736.64 1.00x asm: benchmark old MB/s new MB/s speedup BenchmarkHashes/xxhash,n=5_B-4 539.76 711.11 1.32x BenchmarkHashes/xxhash,n=100_B-4 4227.89 4598.95 1.09x BenchmarkHashes/xxhash,n=4_KB-4 10590.16 10549.72 1.00x BenchmarkHashes/xxhash,n=10_MB-4 9058.39 9024.04 1.00x
xxhash is a Go implementation of the 64-bit xxHash algorithm, XXH64. This is a high-quality hashing algorithm that is much faster than anything in the Go standard library.
The API is very small, taking its cue from the other hashing packages in the standard library:
$ go doc github.com/cespare/xxhash ! package xxhash // import "github.com/cespare/xxhash" Package xxhash implements the 64-bit variant of xxHash (XXH64) as described at http://cyan4973.github.io/xxHash/. func New() hash.Hash64 func Sum64(b []byte) uint64
This implementation provides a fast pure-Go implementation and an even faster assembly implementation for amd64.
Here are some quick benchmarks comparing the pure-Go and assembly implementations of Sum64 against another popular Go XXH64 implementation, github.com/OneOfOne/xxhash:
| input size | OneOfOne | cespare (noasm) | cespare |
|---|---|---|---|
| 5 B | 438.34 MB/s | 596.40 MB/s | 711.11 MB/s |
| 100 B | 3676.54 MB/s | 4301.40 MB/s | 4598.95 MB/s |
| 4 KB | 8128.64 MB/s | 8840.83 MB/s | 10549.72 MB/s |
| 10 MB | 7335.19 MB/s | 7736.64 MB/s | 9024.04 MB/s |