// Copyright 2011 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Large data benchmark.
// The JSON data is a summary of agl's changes in the
// go, webkit, and chromium open source projects.
// We benchmark converting between the JSON form
// and in-memory data structures.
package json
import (
"bytes"
"compress/gzip"
"fmt"
"internal/testenv"
"io/ioutil"
"os"
"reflect"
"runtime"
"strings"
"sync"
"testing"
)
type codeResponse struct {
Tree *codeNode `json:"tree"`
Username string `json:"username"`
}
type codeNode struct {
Name string `json:"name"`
Kids []*codeNode `json:"kids"`
CLWeight float64 `json:"cl_weight"`
Touches int `json:"touches"`
MinT int64 `json:"min_t"`
MaxT int64 `json:"max_t"`
MeanT int64 `json:"mean_t"`
}
var codeJSON []byte
var codeStruct codeResponse
func codeInit() {
f, err := os.Open("testdata/code.json.gz")
if err != nil {
panic(err)
}
defer f.Close()
gz, err := gzip.NewReader(f)
if err != nil {
panic(err)
}
data, err := ioutil.ReadAll(gz)
if err != nil {
panic(err)
}
codeJSON = data
if err := Unmarshal(codeJSON, &codeStruct); err != nil {
panic("unmarshal code.json: " + err.Error())
}
if data, err = Marshal(&codeStruct); err != nil {
panic("marshal code.json: " + err.Error())
}
if !bytes.Equal(data, codeJSON) {
println("different lengths", len(data), len(codeJSON))
for i := 0; i < len(data) && i < len(codeJSON); i++ {
if data[i] != codeJSON[i] {
println("re-marshal: changed at byte", i)
println("orig: ", string(codeJSON[i-10:i+10]))
println("new: ", string(data[i-10:i+10]))
break
}
}
panic("re-marshal code.json: different result")
}
}
func BenchmarkCodeEncoder(b *testing.B) {
if codeJSON == nil {
b.StopTimer()
codeInit()
b.StartTimer()
}
b.RunParallel(func(pb *testing.PB) {
enc := NewEncoder(ioutil.Discard)
for pb.Next() {
if err := enc.Encode(&codeStruct); err != nil {
b.Fatal("Encode:", err)
}
}
})
b.SetBytes(int64(len(codeJSON)))
}
func BenchmarkCodeMarshal(b *testing.B) {
if codeJSON == nil {
b.StopTimer()
codeInit()
b.StartTimer()
}
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
if _, err := Marshal(&codeStruct); err != nil {
b.Fatal("Marshal:", err)
}
}
})
b.SetBytes(int64(len(codeJSON)))
}
func benchMarshalBytes(n int) func(*testing.B) {
sample := []byte("hello world")
// Use a struct pointer, to avoid an allocation when passing it as an
// interface parameter to Marshal.
v := &struct {
Bytes []byte
}{
bytes.Repeat(sample, (n/len(sample))+1)[:n],
}
return func(b *testing.B) {
for i := 0; i < b.N; i++ {
if _, err := Marshal(v); err != nil {
b.Fatal("Marshal:", err)
}
}
}
}
func BenchmarkMarshalBytes(b *testing.B) {
// 32 fits within encodeState.scratch.
b.Run("32", benchMarshalBytes(32))
// 256 doesn't fit in encodeState.scratch, but is small enough to
// allocate and avoid the slower base64.NewEncoder.
b.Run("256", benchMarshalBytes(256))
// 4096 is large enough that we want to avoid allocating for it.
b.Run("4096", benchMarshalBytes(4096))
}
func BenchmarkCodeDecoder(b *testing.B) {
if codeJSON == nil {
b.StopTimer()
codeInit()
b.StartTimer()
}
b.RunParallel(func(pb *testing.PB) {
var buf bytes.Buffer
dec := NewDecoder(&buf)
var r codeResponse
for pb.Next() {
buf.Write(codeJSON)
// hide EOF
buf.WriteByte('\n')
buf.WriteByte('\n')
buf.WriteByte('\n')
if err := dec.Decode(&r); err != nil {
b.Fatal("Decode:", err)
}
}
})
b.SetBytes(int64(len(codeJSON)))
}
func BenchmarkUnicodeDecoder(b *testing.B) {
j := []byte(`"\uD83D\uDE01"`)
b.SetBytes(int64(len(j)))
r := bytes.NewReader(j)
dec := NewDecoder(r)
var out string
b.ResetTimer()
for i := 0; i < b.N; i++ {
if err := dec.Decode(&out); err != nil {
b.Fatal("Decode:", err)
}
r.Seek(0, 0)
}
}
func BenchmarkDecoderStream(b *testing.B) {
b.StopTimer()
var buf bytes.Buffer
dec := NewDecoder(&buf)
buf.WriteString(`"` + strings.Repeat("x", 1000000) + `"` + "\n\n\n")
var x interface{}
if err := dec.Decode(&x); err != nil {
b.Fatal("Decode:", err)
}
ones := strings.Repeat(" 1\n", 300000) + "\n\n\n"
b.StartTimer()
for i := 0; i < b.N; i++ {
if i%300000 == 0 {
buf.WriteString(ones)
}
x = nil
if err := dec.Decode(&x); err != nil || x != 1.0 {
b.Fatalf("Decode: %v after %d", err, i)
}
}
}
func BenchmarkCodeUnmarshal(b *testing.B) {
if codeJSON == nil {
b.StopTimer()
codeInit()
b.StartTimer()
}
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
var r codeResponse
if err := Unmarshal(codeJSON, &r); err != nil {
b.Fatal("Unmarshal:", err)
}
}
})
b.SetBytes(int64(len(codeJSON)))
}
func BenchmarkCodeUnmarshalReuse(b *testing.B) {
if codeJSON == nil {
b.StopTimer()
codeInit()
b.StartTimer()
}
b.RunParallel(func(pb *testing.PB) {
var r codeResponse
for pb.Next() {
if err := Unmarshal(codeJSON, &r); err != nil {
b.Fatal("Unmarshal:", err)
}
}
})
// TODO(bcmills): Is there a missing b.SetBytes here?
}
func BenchmarkUnmarshalString(b *testing.B) {
data := []byte(`"hello, world"`)
b.RunParallel(func(pb *testing.PB) {
var s string
for pb.Next() {
if err := Unmarshal(data, &s); err != nil {
b.Fatal("Unmarshal:", err)
}
}
})
}
func BenchmarkUnmarshalFloat64(b *testing.B) {
data := []byte(`3.14`)
b.RunParallel(func(pb *testing.PB) {
var f float64
for pb.Next() {
if err := Unmarshal(data, &f); err != nil {
b.Fatal("Unmarshal:", err)
}
}
})
}
func BenchmarkUnmarshalInt64(b *testing.B) {
data := []byte(`3`)
b.RunParallel(func(pb *testing.PB) {
var x int64
for pb.Next() {
if err := Unmarshal(data, &x); err != nil {
b.Fatal("Unmarshal:", err)
}
}
})
}
func BenchmarkIssue10335(b *testing.B) {
b.ReportAllocs()
j := []byte(`{"a":{ }}`)
b.RunParallel(func(pb *testing.PB) {
var s struct{}
for pb.Next() {
if err := Unmarshal(j, &s); err != nil {
b.Fatal(err)
}
}
})
}
func BenchmarkUnmapped(b *testing.B) {
b.ReportAllocs()
j := []byte(`{"s": "hello", "y": 2, "o": {"x": 0}, "a": [1, 99, {"x": 1}]}`)
b.RunParallel(func(pb *testing.PB) {
var s struct{}
for pb.Next() {
if err := Unmarshal(j, &s); err != nil {
b.Fatal(err)
}
}
})
}
func BenchmarkTypeFieldsCache(b *testing.B) {
var maxTypes int = 1e6
if testenv.Builder() != "" {
maxTypes = 1e3 // restrict cache sizes on builders
}
// Dynamically generate many new types.
types := make([]reflect.Type, maxTypes)
fs := []reflect.StructField{{
Type: reflect.TypeOf(""),
Index: []int{0},
}}
for i := range types {
fs[0].Name = fmt.Sprintf("TypeFieldsCache%d", i)
types[i] = reflect.StructOf(fs)
}
// clearClear clears the cache. Other JSON operations, must not be running.
clearCache := func() {
fieldCache = sync.Map{}
}
// MissTypes tests the performance of repeated cache misses.
// This measures the time to rebuild a cache of size nt.
for nt := 1; nt <= maxTypes; nt *= 10 {
ts := types[:nt]
b.Run(fmt.Sprintf("MissTypes%d", nt), func(b *testing.B) {
nc := runtime.GOMAXPROCS(0)
for i := 0; i < b.N; i++ {
clearCache()
var wg sync.WaitGroup
for j := 0; j < nc; j++ {
wg.Add(1)
go func(j int) {
for _, t := range ts[(j*len(ts))/nc : ((j+1)*len(ts))/nc] {
cachedTypeFields(t)
}
wg.Done()
}(j)
}
wg.Wait()
}
})
}
// HitTypes tests the performance of repeated cache hits.
// This measures the average time of each cache lookup.
for nt := 1; nt <= maxTypes; nt *= 10 {
// Pre-warm a cache of size nt.
clearCache()
for _, t := range types[:nt] {
cachedTypeFields(t)
}
b.Run(fmt.Sprintf("HitTypes%d", nt), func(b *testing.B) {
b.RunParallel(func(pb *testing.PB) {
for pb.Next() {
cachedTypeFields(types[0])
}
})
})
}
}