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
|
/********************************************************************************
* *
* W o r k - S t e a l i n g Q u e u e *
* *
*********************************************************************************
* Copyright (C) 2016,2022 by Jeroen van der Zijp. All Rights Reserved. *
*********************************************************************************
* This library is free software; you can redistribute it and/or modify *
* it under the terms of the GNU Lesser General Public License as published by *
* the Free Software Foundation; either version 3 of the License, or *
* (at your option) any later version. *
* *
* This library is distributed in the hope that it will be useful, *
* but WITHOUT ANY WARRANTY; without even the implied warranty of *
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the *
* GNU Lesser General Public License for more details. *
* *
* You should have received a copy of the GNU Lesser General Public License *
* along with this program. If not, see <http://www.gnu.org/licenses/> *
********************************************************************************/
#include "xincs.h"
#include "fxdefs.h"
#include "fxmath.h"
#include "FXElement.h"
#include "FXArray.h"
#include "FXPtrList.h"
#include "FXAtomic.h"
#include "FXWSQueue.h"
/*
Notes:
- A SPMC (Single Producer, Multiple Consumer) queue that minimizes contention
of shared state.
- Only a single thread may push() and pop() at the front of the queue, but multiple
threads may take() (steal) items of the back of the queue.
- It is lock-free in that no operating-system calls are used; this solution requires
atomic operations however.
- This is a version of the Chase-Lev work-stealing queue.
*/
using namespace FX;
/*******************************************************************************/
namespace FX {
// Create a queue and set its size to sz
FXWSQueue::FXWSQueue(FXint sz):top(0),bot(0){
setSize(sz);
}
// Change size of the queue
FXbool FXWSQueue::setSize(FXint sz){
if(sz&(sz-1)){ fxerror("FXWSQueue::setSize: bad argument: %d.\n",sz); }
if(list.no(sz)){
top=bot=0;
return true;
}
return false;
}
// Return used slots
FXint FXWSQueue::getUsed() const {
return bot-top;
}
// Return free slots
FXint FXWSQueue::getFree() const {
return getSize()-bot+top;
}
// Check if queue is full
FXbool FXWSQueue::isFull() const {
return (bot-top)>=getSize();
}
// Check if queue is empty
FXbool FXWSQueue::isEmpty() const {
return (bot-top)<=0;
}
// Push task
FXbool FXWSQueue::push(FXptr ptr){
FXint mask=getSize()-1;
FXint t=top;
if(__likely((bot-t)<mask)){
list[bot&mask]=ptr;
bot++;
atomicThreadFence();
return true;
}
return false;
}
// Pop task
FXbool FXWSQueue::pop(FXptr& ptr){
FXint mask=getSize()-1;
FXint b=bot-1;
bot=b;
atomicThreadFence();
FXival t=top;
if(t<=b){
ptr=list[b&mask];
if(t==b){
if(!atomicBoolCas(&top,t,t+1)){
ptr=nullptr;
bot=b+1;
atomicThreadFence();
return false;
}
bot=b+1;
atomicThreadFence();
}
return true;
}
ptr=nullptr;
bot=b+1;
atomicThreadFence();
return false;
}
// Take (steal) task
FXbool FXWSQueue::take(FXptr& ptr){
FXint mask=getSize()-1;
FXint t=top;
atomicThreadFence();
FXint b=bot;
if(__likely(t<b)){
ptr=list[t&mask];
if(atomicBoolCas(&top,t,t+1)){
return true;
}
}
ptr=nullptr;
return false;
}
// Delete queue
FXWSQueue::~FXWSQueue(){
}
}
|
