hpcbits/LFMPMCQueue.h at main · ik80/hpcbits · GitHub

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
152
153
154
155
156
157
/*
 * LFMPSCQueue.h
 *
 * Code borrowed from blogpost by Dmitry Vyukov: http://www.1024cores.net/home/lock-free-algorithms/queues/bounded-mpmc-queue
 *
 */

#ifndef LFMPSCQUEUE_H_
#define LFMPSCQUEUE_H_

#include <atomic>
#include <exception>

// TODO: figure out proper noexcept operator syntax to properly apply noexcept to push/pop/moveTo based on T
template<typename T>
class LFMPMCQueue
{
public:
    LFMPMCQueue(size_t maxQueueSize) :
                    elements(new Element[maxQueueSize]), maxQueueSizeMask(maxQueueSize - 1)
    {
        if(!((maxQueueSize >= 2) && ((maxQueueSize & (maxQueueSize - 1)) == 0)))
            abort();

        for(size_t i = 0; i != maxQueueSize; i += 1)
            elements[i].sequence.store(i, std::memory_order_relaxed);
        head.store(0, std::memory_order_relaxed);
        tail.store(0, std::memory_order_relaxed);
    }

    ~LFMPMCQueue()
    {
        delete[] elements;
    }

    bool push(T const& data)
    {
        Element* pElement;
        size_t pos = head.load(std::memory_order_relaxed);
        for(;;)
        {
            pElement = &elements[pos & maxQueueSizeMask];
            size_t seq = pElement->sequence.load(std::memory_order_acquire);
            intptr_t dif = (intptr_t) seq - (intptr_t) pos;
            if(dif == 0)
            {
                if(head.compare_exchange_weak(pos, pos + 1, std::memory_order_relaxed))
                    break;
            }
            else if(dif < 0)
                return false;
            else
                pos = head.load(std::memory_order_relaxed);
        }
        pElement->data = data;
        pElement->sequence.store(pos + 1, std::memory_order_release);
        return true;
    }

    bool moveFrom(T&& data)
    {
        Element* pElement;
        size_t pos = head.load(std::memory_order_relaxed);
        for(;;)
        {
            pElement = &elements[pos & maxQueueSizeMask];
            size_t seq = pElement->sequence.load(std::memory_order_acquire);
            intptr_t dif = (intptr_t) seq - (intptr_t) pos;
            if(dif == 0)
            {
                if(head.compare_exchange_weak(pos, pos + 1, std::memory_order_relaxed))
                    break;
            }
            else if(dif < 0)
                return false;
            else
                pos = head.load(std::memory_order_relaxed);
        }
        pElement->data = std::move(data);
        pElement->sequence.store(pos + 1, std::memory_order_release);
        return true;
    }

    bool pop(T& data)
    {
        Element* pElement;
        size_t pos = tail.load(std::memory_order_relaxed);
        for(;;)
        {
            pElement = &elements[pos & maxQueueSizeMask];
            size_t seq = pElement->sequence.load(std::memory_order_acquire);
            intptr_t dif = (intptr_t) seq - (intptr_t) (pos + 1);
            if(dif == 0)
            {
                if(tail.compare_exchange_weak(pos, pos + 1, std::memory_order_relaxed))
                    break;
            }
            else if(dif < 0)
                return false;
            else
                pos = tail.load(std::memory_order_relaxed);
        }
        data = pElement->data;
        pElement->sequence.store(pos + maxQueueSizeMask + 1, std::memory_order_release);
        return true;
    }

    bool moveTo(T& data)
    {
        Element* pElement;
        size_t pos = tail.load(std::memory_order_relaxed);
        for(;;)
        {
            pElement = &elements[pos & maxQueueSizeMask];
            size_t seq = pElement->sequence.load(std::memory_order_acquire);
            intptr_t dif = (intptr_t) seq - (intptr_t) (pos + 1);
            if(dif == 0)
            {
                if(tail.compare_exchange_weak(pos, pos + 1, std::memory_order_relaxed))
                    break;
            }
            else if(dif < 0)
                return false;
            else
                pos = tail.load(std::memory_order_relaxed);
        }
        data = std::move(pElement->data);
        pElement->sequence.store(pos + maxQueueSizeMask + 1, std::memory_order_release);
        return true;
    }

private:
    struct Element
    {
        std::atomic_size_t sequence;
        T data;
    };

    static const size_t CACHE_LINE_SIZE = 64;

    Element* elements;
    const size_t maxQueueSizeMask;
    char pad0[CACHE_LINE_SIZE - sizeof(size_t) - sizeof(Element*)];
    std::atomic_size_t head;
    char pad1[CACHE_LINE_SIZE - sizeof(std::atomic_size_t)];
    std::atomic_size_t tail;
    char pad2[CACHE_LINE_SIZE - sizeof(std::atomic_size_t)];

    // deleted functions
    LFMPMCQueue(const LFMPMCQueue&);
    void operator=(const LFMPMCQueue&);
    LFMPMCQueue(LFMPMCQueue&&);
    void operator =(LFMPMCQueue&&);

}__attribute((aligned(CACHE_LINE_SIZE)));

#endif /* LFMPSCQUEUE_H_ */