c++ - 哪些 std::async 实现使用线程池？

Question

std::async使用而不是手动创建对象的优点之一std::thread应该是std::async可以在幕后使用线程池来避免超额订阅问题。但是哪些实现可以做到这一点？我的理解是微软的实现可以，但是这些其他的async实现呢？

Gnu 的 libstdc++
LLVM 的 libc++
Just Software的图书馆
提升（对于boost::thread::async，不是std::async）

感谢您提供的任何信息。

score 44 · Accepted Answer

黑盒测试

虽然“白盒”检查可以通过检查 boost、libstdc++ 或 libc++ 源，或检查文档如just::thread或MSVC Concurrency Runtime来完成，但我不能否认自己编写 C++11 代码的乐趣！

我做了黑盒测试：

现场演示

#define BOOST_THREAD_PROVIDES_FUTURE
#define BOOST_RESULT_OF_USE_DECLTYPE
#include <boost/exception/exception.hpp>
#include <boost/range/algorithm.hpp>
#include <boost/move/iterator.hpp>
#include <boost/phoenix.hpp>
#include <boost/thread.hpp>
#include <boost/config.hpp>

#include <unordered_set>
#include <functional>
#include <algorithm>
#include <iterator>
#include <iostream>
#include <ostream>
#include <cstddef>
#include <string>
#include <vector>
#include <future>
#include <thread>
#include <chrono>
#include <mutex>

// _____________________[CONFIGURATION]________________________ //
namespace async_lib = std;
const bool work_is_sleep = false;
// ____________________________________________________________ //

using namespace std;
using boost::phoenix::arg_names::arg1;
using boost::thread_specific_ptr;
using boost::back_move_inserter;
using boost::type_name;
using boost::count_if;
using boost::copy;

template<typename Mutex>
unique_lock<Mutex> locker(Mutex &m)
{
    return unique_lock<Mutex>(m);
}

void do_work()
{
    if(work_is_sleep)
    {
        this_thread::sleep_for( chrono::milliseconds(20) );
    }
    else
    {
        volatile double result=0.0;
        for(size_t i=0; i!=1<<22; ++i)
            result+=0.1;
    }
}

int main()
{
    typedef thread::id TID;
    typedef async_lib::future<TID> FTID;

    unordered_set<TID> tids, live_tids;
    vector<FTID> ftids;
    vector<int> live_tids_count;
    async_lib::mutex m;
    generate_n
    (
        back_move_inserter(ftids), 64*thread::hardware_concurrency(),
        [&]()
        {
            return async_lib::async([&]() -> TID
            {
                static thread_specific_ptr<bool> fresh;
                if(fresh.get() == nullptr)
                    fresh.reset(new bool(true));
                TID tid = this_thread::get_id();
                locker(m),
                    live_tids.insert(tid),
                    live_tids_count.push_back(int(live_tids.size()) * (*fresh ? -1 : 1));
                do_work();
                locker(m),
                    live_tids.erase(tid);
                *fresh = false;
                return tid;
            });
        }
    );
    transform
    (
        begin(ftids), end(ftids),
        inserter(tids, tids.end()),
        [](FTID &x){return x.get();}
    );

    cout << "Compiler = " << BOOST_COMPILER                             << endl;
    cout << "Standard library = " << BOOST_STDLIB                       << endl;
    cout << "Boost = " << BOOST_LIB_VERSION                             << endl;
    cout << "future type = " << type_name<FTID>()                       << endl;
    cout << "Only sleep in do_work = " << boolalpha << work_is_sleep    << endl;
    cout << string(32,'_')                                              << endl;
    cout << "hardware_concurrency = " << thread::hardware_concurrency() << endl;
    cout << "async count = " << ftids.size()                            << endl;
    cout << "unique thread id's = " << tids.size()                      << endl;
    cout << "live threads count (negative means fresh thread):"                ;
    copy(live_tids_count, ostream_iterator<int>(cout," "));        cout << endl;
    cout << "fresh count = " << count_if(live_tids_count, arg1 < 0)     << endl;
}

Compiler = Microsoft Visual C++ version 11.0
Standard library = Dinkumware standard library version 540
Boost = 1_53
future type = class std::future<class std::thread::id>
Only sleep in do_work = false
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 4
live threads count (negative means fresh thread):-1 -2 2 2 -3 -4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4
 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 4 4 4 4 4 4 4 4 4
 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4
 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4
fresh count = 4

Compiler = Microsoft Visual C++ version 11.0
Standard library = Dinkumware standard library version 540
Boost = 1_53
future type = class std::future<class std::thread::id>
Only sleep in do_work = true
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 34
live threads count (negative means fresh thread):-1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 -13 -14 -15 -16 -17 -18 18 18 18 18 18 18 18 18 18 18 18 18 18
 18 18 18 18 -19 19 -20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 20 -21 21 21 -22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 22 -
23 23 23 23 -24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 24 -25 25 25 25 25 -26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26 26
26 26 26 -27 27 27 27 27 27 -28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 28 -29 29 29 29 29 29 29 -30 30 30 30 30 30 30 30 30 30
 30 30 30 30 30 30 30 30 30 30 30 30 30 30 -31 31 31 31 31 31 31 31 -32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 32 -33 33
 33 33 33 33 33 33 33 -34 34 34 34 34 34 34 34 34 11 12 11 12 13 14 15 16 15 10 11 12 13 14
fresh count = 34

Compiler = Microsoft Visual C++ version 11.0
Standard library = Dinkumware standard library version 540
Boost = 1_53
future type = class boost::future<class std::thread::id>
Only sleep in do_work = false
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 256
live threads count (negative means fresh thread):-1 -2 -2 -3 -3 -3 -3 -3 -3 -3 -3 -3 -3 -3 -3 -3 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -3 -4 -4 -4 -4 -4 -4 -4
 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -4 -2 -3 -3 -3 -3 -4 -4 -4 -4 -2 -2 -2 -2 -3
 -2 -2 -3 -1 -2 -2 -3 -3 -1 -2 -1 -2 -3 -2 -2 -2 -2 -1 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -3 -3 -2 -2 -2 -2 -3 -3 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2
 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -3 -1 -2 -1 -2 -1 -2 -1 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -3 -3 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2
 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -3 -3 -2 -2 -3 -1 -2 -1 -2 -2 -2 -3 -2 -3 -1 -2 -2 -2 -3 -2 -3 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2
 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -2 -3 -3 -3 -2 -2 -2 -3
fresh count = 256

Compiler = Microsoft Visual C++ version 11.0
Standard library = Dinkumware standard library version 540
Boost = 1_53
future type = class boost::future<class std::thread::id>
Only sleep in do_work = true
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 256
live threads count (negative means fresh thread):-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 -2
8 -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 -158 -159 -160 -161 -162
 -163 -164 -165 -166 -167 -168 -169 -170 -171 -172 -173 -174 -175 -176 -177 -178 -179 -180 -181 -182 -183 -184 -185 -186 -187 -188 -189 -190 -191 -192
 -193 -194 -195 -196 -197 -198 -199 -200 -201 -202 -203 -204 -205 -206 -207 -208 -209 -210 -211 -212 -213 -214 -215 -216 -217 -218 -219 -220 -221 -222
 -223 -224 -225 -226 -227 -228 -229 -230 -231 -232 -233 -234 -235 -236 -237 -238 -239 -240 -241 -242 -243 -244 -245 -246 -247 -248 -249 -250 -251 -252
 -253 -254 -255 -256
fresh count = 256

Compiler = GNU C++ version 4.8.0-alpha20121216 20121216 (experimental)
Standard library = GNU libstdc++ version 20121216
Boost = 1_53
future type = std::future<std::thread::id>
Only sleep in do_work = false
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 1
live threads count (negative means fresh thread):-1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
fresh count = 1

Compiler = GNU C++ version 4.8.0-alpha20121216 20121216 (experimental)
Standard library = GNU libstdc++ version 20121216
Boost = 1_53
future type = boost::future<std::thread::id>
Only sleep in do_work = false
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 122
live threads count (negative means fresh thread):-1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 -13 -14 -15 -16 -17 -18 -19 -20 -21 -20 -21 -22 -23 -24 -25 -25 -26 -27 -28 -29 -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 -59 -60 -61 -62 -63 -64 -64 -64 -63 -61 -55 -48 -41 -33 -34 -27 -27 -28 -29 -29 -30 -31 -31 -30 -28 -29 -28 -28 -28 -29 -30 -30 -31 -30 -31 -31 -31 -32 -31 -26 -24 -25 -26 -25 -23 -23 -22 -20 -21 -19 -20 -20 -19 -20 -21 -21 -22 -21 -22 -23 -24 -24 -25 -26 -27 -25 -25 -25 -22 -23 -22 -23 -23 -24 -25 -26 -27 -27 -28 -29 -29 -30 -31 -32 -33 -34 -35 -36 -37 -38 -39 -40 -41 -42 -43 -44 -45 -46 -47 -48 -49 -50 -51 -51 -52 -53 -54 -55 -56 -57 -56 -54 -3 -3 -4 -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 -55 -56 -57 -58 -59 -60 -61 -61 -61 -62 -63 -64 -65 -64 -62 -63 -63 -61 -62 -61 -59 -60 -59 -57 -55 -56
fresh count = 256

Compiler = Clang version 3.2 (tags/RELEASE_32/final)
Standard library = libc++ version 1101
Boost = 1_53
future type = NSt3__16futureINS_11__thread_idEEE
Only sleep in do_work = false
________________________________
hardware_concurrency = 4
async count = 256
unique thread id's = 255
live threads count (negative means fresh thread):-1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 -12 -13 -14 -15 -16 -17 -18 -19 -20 -21 -22 -23 -24 -25 -26 -27 -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 -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 -102 -103 -104 -105 -106 -107 -108 -109 -110 -111 -112 -113 -114 -115 -116 -117 -118 -119 -120 -121 -121 -122 -122 -123 -123 -124 -125 -126 -127 -126 -127 -128 -129 -129 -128 -129 -128 -129 -129 -128 -129 -129 -128 -126 -127 -128 -128 -129 -129 -130 -129 -129 -128 -129 -129 -130 -131 -132 -133 -134 -135 -136 -134 -132 -133 -134 -134 -133 -132 -133 -132 -133 -134 -133 -131 -129 -127 -124 -125 -121 -119 -120 -118 -119 -118 -117 -115 -111 -107 -105 -106 -103 -100 -97 -95 -96 -94 -90 -87 -81 -73 -74 -71 -72 -73 -74 -75 -70 -71 -66 -60 -59 -60 -61 -62 -63 -64 -61 -58 -55 -55 -52 -53 -54 -54 -55 -56 -56 -57 -54 -55 -56 -56 -57 -57 -58 -56 -54 -55 -56 -56 -57 -58 -58 -59 -58 -58 -58 -58 -59 -60
fresh count = 256

可以看出——MSVC 确实重用了线程，也就是说，它的方案很可能是某种线程池。

c++ - 哪些 std::async 实现使用线程池？

1 回答 1

黑盒测试

Related

Reference