1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
|
package bytebufferpool
import (
"sort"
"sync"
"sync/atomic"
)
const (
minBitSize = 6 // 2**6=64 is a CPU cache line size
steps = 20
minSize = 1 << minBitSize
maxSize = 1 << (minBitSize + steps - 1)
calibrateCallsThreshold = 42000
maxPercentile = 0.95
)
// Pool represents byte buffer pool.
//
// Distinct pools may be used for distinct types of byte buffers.
// Properly determined byte buffer types with their own pools may help reducing
// memory waste.
type Pool struct {
calls [steps]uint64
calibrating uint64
defaultSize uint64
maxSize uint64
pool sync.Pool
}
var defaultPool Pool
// Get returns an empty byte buffer from the pool.
//
// Got byte buffer may be returned to the pool via Put call.
// This reduces the number of memory allocations required for byte buffer
// management.
func Get() *ByteBuffer { return defaultPool.Get() }
// Get returns new byte buffer with zero length.
//
// The byte buffer may be returned to the pool via Put after the use
// in order to minimize GC overhead.
func (p *Pool) Get() *ByteBuffer {
v := p.pool.Get()
if v != nil {
return v.(*ByteBuffer)
}
return &ByteBuffer{
B: make([]byte, 0, atomic.LoadUint64(&p.defaultSize)),
}
}
// Put returns byte buffer to the pool.
//
// ByteBuffer.B mustn't be touched after returning it to the pool.
// Otherwise data races will occur.
func Put(b *ByteBuffer) { defaultPool.Put(b) }
// Put releases byte buffer obtained via Get to the pool.
//
// The buffer mustn't be accessed after returning to the pool.
func (p *Pool) Put(b *ByteBuffer) {
idx := index(len(b.B))
if atomic.AddUint64(&p.calls[idx], 1) > calibrateCallsThreshold {
p.calibrate()
}
maxSize := int(atomic.LoadUint64(&p.maxSize))
if maxSize == 0 || cap(b.B) <= maxSize {
b.Reset()
p.pool.Put(b)
}
}
func (p *Pool) calibrate() {
if !atomic.CompareAndSwapUint64(&p.calibrating, 0, 1) {
return
}
a := make(callSizes, 0, steps)
var callsSum uint64
for i := uint64(0); i < steps; i++ {
calls := atomic.SwapUint64(&p.calls[i], 0)
callsSum += calls
a = append(a, callSize{
calls: calls,
size: minSize << i,
})
}
sort.Sort(a)
defaultSize := a[0].size
maxSize := defaultSize
maxSum := uint64(float64(callsSum) * maxPercentile)
callsSum = 0
for i := 0; i < steps; i++ {
if callsSum > maxSum {
break
}
callsSum += a[i].calls
size := a[i].size
if size > maxSize {
maxSize = size
}
}
atomic.StoreUint64(&p.defaultSize, defaultSize)
atomic.StoreUint64(&p.maxSize, maxSize)
atomic.StoreUint64(&p.calibrating, 0)
}
type callSize struct {
calls uint64
size uint64
}
type callSizes []callSize
func (ci callSizes) Len() int {
return len(ci)
}
func (ci callSizes) Less(i, j int) bool {
return ci[i].calls > ci[j].calls
}
func (ci callSizes) Swap(i, j int) {
ci[i], ci[j] = ci[j], ci[i]
}
func index(n int) int {
n--
n >>= minBitSize
idx := 0
for n > 0 {
n >>= 1
idx++
}
if idx >= steps {
idx = steps - 1
}
return idx
}
|
