potok/thread-pool
1
0
Fork 0
mirror of https://github.com/bshoshany/thread-pool.git synced 2026-07-21 19:13:00 +04:00
Code Packages Releases Wiki Activity

Updated to v3.0.0

Browse Source
This commit is contained in:
Barak Shoshany
2022-05-31 00:28:22 -04:00
parent 737c603610
commit 9d43f5d05d
13 changed files with 2010 additions and 2042 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/ISSUE_TEMPLATE/bug_report.md
+37
View File
@@ -0,0 +1,37 @@
---
name: Bug report
about: Found a bug? Report it here.
title: "[BUG]"
labels: bug
assignees: bshoshany
---
**Describe the bug**
A clear and concise description of what the bug is.
**Minimal working example**
A short but complete program that can be compiled to reproduce the error. Paste the program between the two code fences. If it's too long or requires multiple files, attach the file(s) instead.
```cpp
```
**Behavior**
What behavior did you expect to get? What actually happened? If the code failed to compile, please include the full output of the compiler.
**System information**
* CPU model, architecture, # of cores and threads:
* Operating system:
* Name and version of C++ compiler:
* Full command used for compiling, including all compiler flags:
* Thread pool library version:
(Please note that only the latest version of the thread pool library is supported.)
**Additional information**
Include any additional information here.
.github/ISSUE_TEMPLATE/failed-tests.md
+26
View File
@@ -0,0 +1,26 @@
---
name: Failed tests
about: The provided automated tests failed on your system? Report it here.
title: "[TEST]"
labels: bug
assignees: bshoshany
---
**System information**
* CPU model, architecture, # of cores and threads:
* Operating system:
* Name and version of C++ compiler:
* Full command used for compiling, including all compiler flags:
* Thread pool library version:
(Please note that only the latest version of the thread pool library is supported.)
**Log file**
Please attach the log file generated by the automated test program to this issue.
**Additional information**
Include any additional information here.
.github/ISSUE_TEMPLATE/feature_request.md
+23
View File
@@ -0,0 +1,23 @@
---
name: Feature request
about: Want a new feature? Suggest it here.
title: "[REQ]"
labels: enhancement
assignees: bshoshany
---
**Describe the new feature**
A clear and concise description of the feature you want.
**Code example**
An example of code that utilizes the suggested feature. Paste or write it between the two code fences.
```cpp
```
**Additional information**
Include any additional information here.
.github/pull_request_template.md
+20
View File
@@ -0,0 +1,20 @@
**Pull request policy (please read)**
> Contributions are always welcome. However, I release my projects in cumulative updates after editing and testing them locally on my system, so my policy is not to accept any pull requests. If you open a pull request, and I decide to incorporate your suggestion into the project, I will first modify your code to comply with the project's coding conventions (formatting, syntax, naming, comments, programming practices, etc.), and perform some tests to ensure that the change doesn't break anything. I will then merge it into the next release of the project, possibly together with some other changes. The new release will also include a note in `CHANGELOG.md` with a link to your pull request, and modifications to the documentation in `README.md` as needed.
**Describe the changes**
What does your pull request fix or add to the library?
**Testing**
Have you tested the new code using the provided automated test program and/or performed any other tests to ensure that it works correctly? If so, please provide information about the test system(s):
* CPU model, architecture, # of cores and threads:
* Operating system:
* Name and version of C++ compiler:
* Full command used for compiling, including all compiler flags:
**Additional information**
Include any additional information here.
BS_thread_pool.hpp
+516
View File
@@ -0,0 +1,516 @@
#pragma once
/**
* @file BS_thread_pool.hpp
* @author Barak Shoshany (baraksh@gmail.com) (http://baraksh.com)
* @version 3.0.0
* @date 2022-05-30
* @copyright Copyright (c) 2022 Barak Shoshany. Licensed under the MIT license. If you use this library in software of any kind, please provide a link to the GitHub repository https://github.com/bshoshany/thread-pool in the source code and documentation. If you use this library in published research, please cite it as follows: Barak Shoshany, "A C++17 Thread Pool for High-Performance Scientific Computing", doi:10.5281/zenodo.4742687, arXiv:2105.00613 (May 2021)
*
* @brief BS::thread_pool: a fast, lightweight, and easy-to-use C++17 thread pool library. This header file contains the entire library, including the main BS::thread_pool class and the helper classes BS::multi_future, BS:synced_stream, and BS::timer.
*/
#define BS_THREAD_POOL_VERSION "v3.0.0 (2022-05-30)"
#include <atomic> // std::atomic
#include <chrono> // std::chrono
#include <condition_variable> // std::condition_variable
#include <exception> // std::current_exception
#include <functional> // std::function
#include <future> // std::future, std::promise
#include <iostream> // std::cout, std::ostream
#include <memory> // std::make_shared, std::make_unique, std::shared_ptr, std::unique_ptr
#include <mutex> // std::mutex, std::scoped_lock, std::unique_lock
#include <queue> // std::queue
#include <thread> // std::thread
#include <type_traits> // std::common_type_t, std::decay_t, std::is_void_v, std::invoke_result_t
#include <utility> // std::move, std::swap
#include <vector> // std::vector
namespace BS
{
using concurrency_t = std::invoke_result_t<decltype(std::thread::hardware_concurrency)>;
// ============================================================================================= //
// Begin class multi_future //
/**
* @brief A helper class to facilitate waiting for and/or getting the results of multiple futures at once.
*/
template <typename T>
class multi_future
{
public:
/**
* @brief Construct a multi_future object with the given number of futures.
*
* @param num_futures_ The desired number of futures to store.
*/
explicit multi_future(const size_t num_futures_ = 0) : f(num_futures_) {}
/**
* @brief Get the results from all the futures stored in this multi_future object.
*
* @return A vector containing the results.
*/
std::vector<T> get()
{
std::vector<T> results(f.size());
for (size_t i = 0; i < f.size(); ++i)
results[i] = f[i].get();
return results;
}
/**
* @brief Wait for all the futures stored in this multi_future object.
*/
void wait() const
{
for (size_t i = 0; i < f.size(); ++i)
f[i].wait();
}
/**
* @brief A vector to store the futures.
*/
std::vector<std::future<T>> f;
};
// End class multi_future //
// ============================================================================================= //
// ============================================================================================= //
// Begin class thread_pool //
/**
* @brief A fast, lightweight, and easy-to-use C++17 thread pool class.
*/
class thread_pool
{
public:
// ============================
// Constructors and destructors
// ============================
/**
* @brief Construct a new thread pool.
*
* @param thread_count_ The number of threads to use. The default value is the total number of hardware threads available, as reported by the implementation. This is usually determined by the number of cores in the CPU. If a core is hyperthreaded, it will count as two threads.
*/
explicit thread_pool(const concurrency_t thread_count_ = std::thread::hardware_concurrency()) : thread_count(thread_count_ ? thread_count_ : std::thread::hardware_concurrency()), threads(std::make_unique<std::thread[]>(thread_count_ ? thread_count_ : std::thread::hardware_concurrency()))
{
create_threads();
}
/**
* @brief Destruct the thread pool. Waits for all tasks to complete, then destroys all threads. Note that if the variable paused is set to true, then any tasks still in the queue will never be executed.
*/
~thread_pool()
{
wait_for_tasks();
destroy_threads();
}
// =======================
// Public member functions
// =======================
/**
* @brief Get the number of tasks currently waiting in the queue to be executed by the threads.
*
* @return The number of queued tasks.
*/
size_t get_tasks_queued() const
{
const std::scoped_lock tasks_lock(tasks_mutex);
return tasks.size();
}
/**
* @brief Get the number of tasks currently being executed by the threads.
*
* @return The number of running tasks.
*/
size_t get_tasks_running() const
{
const std::scoped_lock tasks_lock(tasks_mutex);
return tasks_total - tasks.size();
}
/**
* @brief Get the total number of unfinished tasks: either still in the queue, or running in a thread. Note that get_tasks_total() == get_tasks_queued() + get_tasks_running().
*
* @return The total number of tasks.
*/
size_t get_tasks_total() const
{
return tasks_total;
}
/**
* @brief Get the number of threads in the pool.
*
* @return The number of threads.
*/
concurrency_t get_thread_count() const
{
return thread_count;
}
/**
* @brief Parallelize a loop by automatically splitting it into blocks and submitting each block separately to the queue.
*
* @tparam F The type of the function to loop through.
* @tparam T1 The type of the first index in the loop. Should be a signed or unsigned integer.
* @tparam T2 The type of the index after the last index in the loop. Should be a signed or unsigned integer. If T1 is not the same as T2, a common type will be automatically inferred.
* @tparam T The common type of T1 and T2.
* @tparam R The return value of the loop function F (can be void).
* @param first_index The first index in the loop.
* @param index_after_last The index after the last index in the loop. The loop will iterate from first_index to (index_after_last - 1) inclusive. In other words, it will be equivalent to "for (T i = first_index; i < index_after_last; ++i)". Note that if first_index == index_after_last, no blocks will be submitted.
* @param loop The function to loop through. Will be called once per block. Should take exactly two arguments: the first index in the block and the index after the last index in the block. loop(start, end) should typically involve a loop of the form "for (T i = start; i < end; ++i)".
* @param num_blocks The maximum number of blocks to split the loop into. The default is to use the number of threads in the pool.
* @return A multi_future object that can be used to wait for all the blocks to finish. If the loop function returns a value, the multi_future object can be used to obtain the values returned by each block.
*/
template <typename F, typename T1, typename T2, typename T = std::common_type_t<T1, T2>, typename R = std::invoke_result_t<std::decay_t<F>, T, T>>
multi_future<R> parallelize_loop(const T1& first_index, const T2& index_after_last, const F& loop, size_t num_blocks = 0)
{
T first_index_T = static_cast<T>(first_index);
T index_after_last_T = static_cast<T>(index_after_last);
if (first_index_T == index_after_last_T)
return multi_future<R>();
if (index_after_last_T < first_index_T)
std::swap(index_after_last_T, first_index_T);
if (num_blocks == 0)
num_blocks = thread_count;
const size_t total_size = static_cast<size_t>(index_after_last_T - first_index_T);
size_t block_size = static_cast<size_t>(total_size / num_blocks);
if (block_size == 0)
{
block_size = 1;
num_blocks = total_size > 1 ? total_size : 1;
}
multi_future<R> mf(num_blocks);
for (size_t i = 0; i < num_blocks; ++i)
{
const T start = (static_cast<T>(i * block_size) + first_index_T);
const T end = (i == num_blocks - 1) ? index_after_last_T : (static_cast<T>((i + 1) * block_size) + first_index_T);
mf.f[i] = submit(loop, start, end);
}
return mf;
}
/**
* @brief Push a function with zero or more arguments, but no return value, into the task queue.
*
* @tparam F The type of the function.
* @tparam A The types of the arguments.
* @param task The function to push.
* @param args The arguments to pass to the function.
*/
template <typename F, typename... A>
void push_task(const F& task, const A&... args)
{
{
const std::scoped_lock tasks_lock(tasks_mutex);
if constexpr (sizeof...(args) == 0)
tasks.push(std::function<void()>(task));
else
tasks.push(std::function<void()>([task, args...] { task(args...); }));
}
++tasks_total;
task_available_cv.notify_one();
}
/**
* @brief Reset the number of threads in the pool. Waits for all currently running tasks to be completed, then destroys all threads in the pool and creates a new thread pool with the new number of threads. Any tasks that were waiting in the queue before the pool was reset will then be executed by the new threads. If the pool was paused before resetting it, the new pool will be paused as well.
*
* @param thread_count_ The number of threads to use. The default value is the total number of hardware threads available, as reported by the implementation. This is usually determined by the number of cores in the CPU. If a core is hyperthreaded, it will count as two threads.
*/
void reset(const concurrency_t thread_count_ = std::thread::hardware_concurrency())
{
const bool was_paused = paused;
paused = true;
wait_for_tasks();
destroy_threads();
thread_count = thread_count_ ? thread_count_ : std::thread::hardware_concurrency();
threads = std::make_unique<std::thread[]>(thread_count);
paused = was_paused;
create_threads();
}
/**
* @brief Submit a function with zero or more arguments into the task queue. If the function has a return value, get a future for the eventual returned value. If the function has no return value, get an std::future<void> which can be used to wait until the task finishes.
*
* @tparam F The type of the function.
* @tparam A The types of the zero or more arguments to pass to the function.
* @tparam R The return type of the function (can be void).
* @param task The function to submit.
* @param args The zero or more arguments to pass to the function.
* @return A future to be used later to wait for the function to finish executing and/or obtain its returned value if it has one.
*/
template <typename F, typename... A, typename R = std::invoke_result_t<std::decay_t<F>, std::decay_t<A>...>>
std::future<R> submit(const F& task, const A&... args)
{
std::shared_ptr<std::promise<R>> task_promise = std::make_shared<std::promise<R>>();
push_task(
[task, args..., task_promise]
{
try
{
if constexpr (std::is_void_v<R>)
{
task(args...);
task_promise->set_value();
}
else
{
task_promise->set_value(task(args...));
}
}
catch (...)
{
try
{
task_promise->set_exception(std::current_exception());
}
catch (...)
{
}
}
});
return task_promise->get_future();
}
/**
* @brief Wait for tasks to be completed. Normally, this function waits for all tasks, both those that are currently running in the threads and those that are still waiting in the queue. However, if the pool is paused, this function only waits for the currently running tasks (otherwise it would wait forever). Note: To wait for just one specific task, use submit() instead, and call the wait() member function of the generated future.
*/
void wait_for_tasks()
{
waiting = true;
std::unique_lock<std::mutex> tasks_lock(tasks_mutex);
task_done_cv.wait(tasks_lock, [this] { return (tasks_total == (paused ? tasks.size() : 0)); });
waiting = false;
}
// ===========
// Public data
// ===========
/**
* @brief An atomic variable indicating whether the workers should pause. When set to true, the workers temporarily stop retrieving new tasks out of the queue, although any tasks already executed will keep running until they are finished. Set to false again to resume retrieving tasks.
*/
std::atomic<bool> paused = false;
private:
// ========================
// Private member functions
// ========================
/**
* @brief Create the threads in the pool and assign a worker to each thread.
*/
void create_threads()
{
running = true;
for (concurrency_t i = 0; i < thread_count; ++i)
{
threads[i] = std::thread(&thread_pool::worker, this);
}
}
/**
* @brief Destroy the threads in the pool.
*/
void destroy_threads()
{
running = false;
task_available_cv.notify_all();
for (concurrency_t i = 0; i < thread_count; ++i)
{
threads[i].join();
}
}
/**
* @brief A worker function to be assigned to each thread in the pool. Waits until it is notified by push_task() that a task is available, and then retrieves the task from the queue and executes it. Once the task finishes, the worker notifies wait_for_tasks() in case it is waiting.
*/
void worker()
{
while (running)
{
std::function<void()> task;
std::unique_lock<std::mutex> tasks_lock(tasks_mutex);
task_available_cv.wait(tasks_lock, [&] { return !tasks.empty() || !running; });
if (running && !paused)
{
task = std::move(tasks.front());
tasks.pop();
tasks_lock.unlock();
task();
--tasks_total;
if (waiting)
task_done_cv.notify_one();
}
}
}
// ============
// Private data
// ============
/**
* @brief An atomic variable indicating to the workers to keep running. When set to false, the workers permanently stop working.
*/
std::atomic<bool> running = false;
/**
* @brief A condition variable used to notify worker() that a new task has become available.
*/
std::condition_variable task_available_cv = {};
/**
* @brief A condition variable used to notify wait_for_tasks() that a tasks is done.
*/
std::condition_variable task_done_cv = {};
/**
* @brief A queue of tasks to be executed by the threads.
*/
std::queue<std::function<void()>> tasks = {};
/**
* @brief An atomic variable to keep track of the total number of unfinished tasks - either still in the queue, or running in a thread.
*/
std::atomic<size_t> tasks_total = 0;
/**
* @brief A mutex to synchronize access to the task queue by different threads.
*/
mutable std::mutex tasks_mutex = {};
/**
* @brief The number of threads in the pool.
*/
concurrency_t thread_count = 0;
/**
* @brief A smart pointer to manage the memory allocated for the threads.
*/
std::unique_ptr<std::thread[]> threads = nullptr;
/**
* @brief An atomic variable indicating that wait_for_tasks() is active and expects to be notified whenever a task is done.
*/
std::atomic<bool> waiting = false;
};
// End class thread_pool //
// ============================================================================================= //
// ============================================================================================= //
// Begin class synced_stream //
/**
* @brief A helper class to synchronize printing to an output stream by different threads.
*/
class synced_stream
{
public:
/**
* @brief Construct a new synced stream.
*
* @param out_stream_ The output stream to print to. The default value is std::cout.
*/
explicit synced_stream(std::ostream& out_stream_ = std::cout) : out_stream(out_stream_) {};
/**
* @brief Print any number of items into the output stream. Ensures that no other threads print to this stream simultaneously, as long as they all exclusively use the same synced_stream object to print.
*
* @tparam T The types of the items
* @param items The items to print.
*/
template <typename... T>
void print(const T&... items)
{
const std::scoped_lock lock(stream_mutex);
(out_stream << ... << items);
}
/**
* @brief Print any number of items into the output stream, followed by a newline character. Ensures that no other threads print to this stream simultaneously, as long as they all exclusively use the same synced_stream object to print.
*
* @tparam T The types of the items
* @param items The items to print.
*/
template <typename... T>
void println(const T&... items)
{
print(items..., '\n');
}
private:
/**
* @brief The output stream to print to.
*/
std::ostream& out_stream;
/**
* @brief A mutex to synchronize printing.
*/
mutable std::mutex stream_mutex = {};
};
// End class synced_stream //
// ============================================================================================= //
// ============================================================================================= //
// Begin class timer //
/**
* @brief A helper class to measure execution time for benchmarking purposes.
*/
class timer
{
public:
/**
* @brief Start (or restart) measuring time.
*/
void start()
{
start_time = std::chrono::steady_clock::now();
}
/**
* @brief Stop measuring time and store the elapsed time since start().
*/
void stop()
{
elapsed_time = std::chrono::steady_clock::now() - start_time;
}
/**
* @brief Get the number of milliseconds that have elapsed between start() and stop().
*
* @return The number of milliseconds.
*/
std::chrono::milliseconds::rep ms() const
{
return (std::chrono::duration_cast<std::chrono::milliseconds>(elapsed_time)).count();
}
private:
/**
* @brief The time point when measuring started.
*/
std::chrono::time_point<std::chrono::steady_clock> start_time = std::chrono::steady_clock::now();
/**
* @brief The duration that has elapsed between start() and stop().
*/
std::chrono::duration<double> elapsed_time = std::chrono::duration<double>::zero();
};
// End class timer //
// ============================================================================================= //
} // namespace BS
BS_thread_pool_test.cpp
+801
View File
@@ -0,0 +1,801 @@
/**
* @file BS_thread_pool_test.cpp
* @author Barak Shoshany (baraksh@gmail.com) (http://baraksh.com)
* @version 3.0.0
* @date 2022-05-30
* @copyright Copyright (c) 2022 Barak Shoshany. Licensed under the MIT license. If you use this library in software of any kind, please provide a link to the GitHub repository https://github.com/bshoshany/thread-pool in the source code and documentation. If you use this library in published research, please cite it as follows: Barak Shoshany, "A C++17 Thread Pool for High-Performance Scientific Computing", doi:10.5281/zenodo.4742687, arXiv:2105.00613 (May 2021)
*
* @brief BS::thread_pool: a fast, lightweight, and easy-to-use C++17 thread pool library. This program tests all aspects of the library, but is not needed in order to use the library.
*/
// Get rid of annoying MSVC warning.
#ifdef _MSC_VER
#define _CRT_SECURE_NO_WARNINGS
#endif
#include <algorithm> // std::min, std::min_element, std::sort, std::unique
#include <atomic> // std::atomic
#include <chrono> // std::chrono
#include <cmath> // std::abs, std::llround, std::round, std::sqrt
#include <ctime> // std::localtime, std::strftime, std::time_t
#include <exception> // std::exception
#include <fstream> // std::ofstream
#include <future> // std::future
#include <iomanip> // std::setprecision, std::setw
#include <ios> // std::fixed
#include <iostream> // std::cout
#include <limits> // std::numeric_limits
#include <random> // std::mt19937_64, std::random_device, std::uniform_int_distribution, std::uniform_real_distribution
#include <stdexcept> // std::runtime_error
#include <string> // std::string, std::to_string
#include <thread> // std::this_thread, std::thread
#include <utility> // std::pair
#include <vector> // std::begin, std::end, std::vector
// Include the header file for the thread pool library.
#include "BS_thread_pool.hpp"
// ================
// Global variables
// ================
// Set to false to disable output to a log file.
constexpr bool output_log = true;
// Set to false to disable testing.
constexpr bool enable_tests = true;
// Set to false to disable the benchmarks.
constexpr bool enable_benchmarks = true;
// Two global synced_streams objects. One prints to std::cout, and the other to a file.
BS::synced_stream sync_cout(std::cout);
std::ofstream log_file;
BS::synced_stream sync_file(log_file);
// A global thread pool object to be used throughout the test.
BS::thread_pool pool;
// A global random_device object to be used to seed some random number generators.
std::random_device rd;
// Global variables to measure how many checks succeeded and how many failed.
size_t tests_succeeded = 0;
size_t tests_failed = 0;
// ================
// Helper functions
// ================
/**
* @brief Print any number of items into both std::cout and the log file, syncing both independently.
*
* @tparam T The types of the items.
* @param items The items to print.
*/
template <typename... T>
void dual_print(const T&... items)
{
sync_cout.print(items...);
if (output_log)
sync_file.print(items...);
}
/**
* @brief Print any number of items into both std::cout and the log file, followed by a newline character, syncing both independently.
*
* @tparam T The types of the items.
* @param items The items to print.
*/
template <typename... T>
void dual_println(const T&... items)
{
dual_print(items..., '\n');
}
/**
* @brief Print a stylized header.
*
* @param text The text of the header. Will appear between two lines.
* @param symbol The symbol to use for the lines. Default is '='.
*/
void print_header(const std::string& text, const char symbol = '=')
{
dual_println();
dual_println(std::string(text.length(), symbol));
dual_println(text);
dual_println(std::string(text.length(), symbol));
}
/**
* @brief Get a string representing the current time.
*
* @return The string.
*/
std::string get_time()
{
const std::time_t t = std::time(nullptr);
char time_string[32];
std::strftime(time_string, sizeof(time_string), "%Y-%m-%d_%H.%M.%S", std::localtime(&t));
return std::string(time_string);
}
/**
* @brief Check if a condition is met, report the result, and keep count of the total number of successes and failures.
*
* @param condition The condition to check.
*/
void check(const bool condition)
{
if (condition)
{
dual_println("-> PASSED!");
++tests_succeeded;
}
else
{
dual_println("-> FAILED!");
++tests_failed;
}
}
// =========================================
// Functions to verify the number of threads
// =========================================
/**
* @brief Store the ID of the current thread in memory. Waits for a short time to ensure it does not get evaluated by more than one thread.
*
* @param location A pointer to the location where the thread ID should be stored.
*/
void store_ID(std::thread::id* location)
{
*location = std::this_thread::get_id();
std::this_thread::sleep_for(std::chrono::milliseconds(10));
}
/**
* @brief Count the number of unique threads in the thread pool to ensure that the correct number of individual threads was created. Pushes a number of tasks equal to four times the thread count into the thread pool, and count the number of unique thread IDs returned by the tasks.
*/
BS::concurrency_t count_unique_threads()
{
std::vector<std::thread::id> thread_IDs(pool.get_thread_count() * 4);
BS::multi_future<void> futures;
for (std::thread::id& id : thread_IDs)
futures.f.push_back(pool.submit(store_ID, &id));
futures.wait();
std::sort(thread_IDs.begin(), thread_IDs.end());
BS::concurrency_t unique_threads = (BS::concurrency_t)(std::unique(thread_IDs.begin(), thread_IDs.end()) - thread_IDs.begin());
return unique_threads;
}
/**
* @brief Check that the constructor works.
*/
void check_constructor()
{
dual_println("Checking that the thread pool reports a number of threads equal to the hardware concurrency...");
check(pool.get_thread_count() == std::thread::hardware_concurrency());
dual_println("Checking that the manually counted number of unique thread IDs is equal to the reported number of threads...");
check(pool.get_thread_count() == count_unique_threads());
}
/**
* @brief Check that reset() works.
*/
void check_reset()
{
pool.reset(std::thread::hardware_concurrency() / 2);
dual_println("Checking that after reset() the thread pool reports a number of threads equal to half the hardware concurrency...");
check(pool.get_thread_count() == std::thread::hardware_concurrency() / 2);
dual_println("Checking that after reset() the manually counted number of unique thread IDs is equal to the reported number of threads...");
check(pool.get_thread_count() == count_unique_threads());
pool.reset(std::thread::hardware_concurrency());
dual_println("Checking that after a second reset() the thread pool reports a number of threads equal to the hardware concurrency...");
check(pool.get_thread_count() == std::thread::hardware_concurrency());
dual_println("Checking that after a second reset() the manually counted number of unique thread IDs is equal to the reported number of threads...");
check(pool.get_thread_count() == count_unique_threads());
}
// =======================================
// Functions to verify submission of tasks
// =======================================
/**
* @brief Check that push_task() works.
*/
void check_push_task()
{
dual_println("Checking that push_task() works for a function with no arguments or return value...");
{
bool flag = false;
pool.push_task([&flag] { flag = true; });
pool.wait_for_tasks();
check(flag);
}
dual_println("Checking that push_task() works for a function with one argument and no return value...");
{
bool flag = false;
pool.push_task([](bool* flag_) { *flag_ = true; }, &flag);
pool.wait_for_tasks();
check(flag);
}
dual_println("Checking that push_task() works for a function with two arguments and no return value...");
{
bool flag1 = false;
bool flag2 = false;
pool.push_task([](bool* flag1_, bool* flag2_) { *flag1_ = *flag2_ = true; }, &flag1, &flag2);
pool.wait_for_tasks();
check(flag1 && flag2);
}
}
/**
* @brief Check that submit() works.
*/
void check_submit()
{
dual_println("Checking that submit() works for a function with no arguments or return value...");
{
bool flag = false;
pool.submit([&flag] { flag = true; }).wait();
check(flag);
}
dual_println("Checking that submit() works for a function with one argument and no return value...");
{
bool flag = false;
pool.submit([](bool* flag_) { *flag_ = true; }, &flag).wait();
check(flag);
}
dual_println("Checking that submit() works for a function with two arguments and no return value...");
{
bool flag1 = false;
bool flag2 = false;
pool.submit([](bool* flag1_, bool* flag2_) { *flag1_ = *flag2_ = true; }, &flag1, &flag2).wait();
check(flag1 && flag2);
}
dual_println("Checking that submit() works for a function with no arguments and a return value...");
{
bool flag = false;
std::future<int> my_future = pool.submit(
[&flag]
{
flag = true;
return 42;
});
check(my_future.get() == 42 && flag);
}
dual_println("Checking that submit() works for a function with one argument and a return value...");
{
bool flag = false;
std::future<int> my_future = pool.submit(
[](bool* flag_)
{
*flag_ = true;
return 42;
},
&flag);
check(my_future.get() == 42 && flag);
}
dual_println("Checking that submit() works for a function with two arguments and a return value...");
{
bool flag1 = false;
bool flag2 = false;
std::future<int> my_future = pool.submit(
[](bool* flag1_, bool* flag2_)
{
*flag1_ = *flag2_ = true;
return 42;
},
&flag1, &flag2);
check(my_future.get() == 42 && flag1 && flag2);
}
}
/**
* @brief Check that wait_for_tasks() works.
*/
void check_wait_for_tasks()
{
const BS::concurrency_t n = pool.get_thread_count() * 10;
std::vector<std::atomic<bool>> flags(n);
for (BS::concurrency_t i = 0; i < n; ++i)
pool.push_task(
[&flags, i]
{
std::this_thread::sleep_for(std::chrono::milliseconds(10));
flags[i] = true;
});
pool.wait_for_tasks();
bool all_flags = true;
for (BS::concurrency_t i = 0; i < n; ++i)
all_flags = all_flags && flags[i];
check(all_flags);
}
// ========================================
// Functions to verify loop parallelization
// ========================================
/**
* @brief Check that parallelize_loop() works for a specific number of indices split over a specific number of tasks, with no return value.
*
* @param start The first index in the loop.
* @param end The last index in the loop plus 1.
* @param num_tasks The number of tasks.
*/
void check_parallelize_loop_no_return(const int64_t random_start, int64_t random_end, const BS::concurrency_t num_tasks)
{
if (random_start == random_end)
++random_end;
dual_println("Verifying that a loop from ", random_start, " to ", random_end, " with ", num_tasks, num_tasks == 1 ? " task" : " tasks", " modifies all indices...");
const size_t num_indices = static_cast<size_t>(std::abs(random_end - random_start));
const int64_t offset = std::min(random_start, random_end);
std::vector<std::atomic<bool>> flags(num_indices);
pool.parallelize_loop(
random_start, random_end,
[&flags, offset](const int64_t start, const int64_t end)
{
for (int64_t i = start; i < end; ++i)
flags[(size_t)(i - offset)] = true;
},
num_tasks)
.wait();
bool all_flags = true;
for (size_t i = 0; i < num_indices; ++i)
all_flags = all_flags && flags[i];
check(all_flags);
}
/**
* @brief Check that parallelize_loop() works for a specific number of indices split over a specific number of tasks, with a return value.
*
* @param start The first index in the loop.
* @param end The last index in the loop plus 1.
* @param num_tasks The number of tasks.
*/
void check_parallelize_loop_return(const int64_t random_start, int64_t random_end, const BS::concurrency_t num_tasks)
{
if (random_start == random_end)
++random_end;
dual_println("Verifying that a loop from ", random_start, " to ", random_end, " with ", num_tasks, num_tasks == 1 ? " task" : " tasks", " correctly sums all indices...");
const std::vector<int64_t> sums_vector = pool.parallelize_loop(
random_start, random_end,
[](const int64_t start, const int64_t end)
{
int64_t total = 0;
for (int64_t i = start; i < end; ++i)
total += i;
return total;
},
num_tasks)
.get();
int64_t sum = 0;
for (const int64_t& s : sums_vector)
sum += s;
check(sum * 2 == std::abs(random_start - random_end) * (random_start + random_end - 1));
}
/**
* @brief Check that parallelize_loop() works using several different random values for the range of indices and number of tasks.
*/
void check_parallelize_loop()
{
std::mt19937_64 mt(rd());
std::uniform_int_distribution<int64_t> index_dist(-1000000, 1000000);
std::uniform_int_distribution<BS::concurrency_t> task_dist(1, pool.get_thread_count());
constexpr uint64_t n = 10;
for (uint64_t i = 0; i < n; ++i)
check_parallelize_loop_no_return(index_dist(mt), index_dist(mt), task_dist(mt));
for (uint64_t i = 0; i < n; ++i)
check_parallelize_loop_return(index_dist(mt), index_dist(mt), task_dist(mt));
}
// ===============================================
// Functions to verify task monitoring and control
// ===============================================
/**
* @brief Check that task monitoring works.
*/
void check_task_monitoring()
{
BS::concurrency_t n = std::min<BS::concurrency_t>(std::thread::hardware_concurrency(), 4);
dual_println("Resetting pool to ", n, " threads.");
pool.reset(n);
dual_println("Submitting ", n * 3, " tasks.");
std::vector<std::atomic<bool>> release(n * 3);
for (BS::concurrency_t i = 0; i < n * 3; ++i)
pool.push_task(
[&release, i]
{
while (!release[i])
std::this_thread::yield();
dual_println("Task ", i, " released.");
});
constexpr std::chrono::milliseconds sleep_time(300);
std::this_thread::sleep_for(sleep_time);
dual_println("After submission, should have: ", n * 3, " tasks total, ", n, " tasks running, ", n * 2, " tasks queued...");
check(pool.get_tasks_total() == n * 3 && pool.get_tasks_running() == n && pool.get_tasks_queued() == n * 2);
for (BS::concurrency_t i = 0; i < n; ++i)
release[i] = true;
std::this_thread::sleep_for(sleep_time);
dual_println("After releasing ", n, " tasks, should have: ", n * 2, " tasks total, ", n, " tasks running, ", n, " tasks queued...");
check(pool.get_tasks_total() == n * 2 && pool.get_tasks_running() == n && pool.get_tasks_queued() == n);
for (BS::concurrency_t i = n; i < n * 2; ++i)
release[i] = true;
std::this_thread::sleep_for(sleep_time);
dual_println("After releasing ", n, " more tasks, should have: ", n, " tasks total, ", n, " tasks running, ", 0, " tasks queued...");
check(pool.get_tasks_total() == n && pool.get_tasks_running() == n && pool.get_tasks_queued() == 0);
for (BS::concurrency_t i = n * 2; i < n * 3; ++i)
release[i] = true;
std::this_thread::sleep_for(sleep_time);
dual_println("After releasing the final ", n, " tasks, should have: ", 0, " tasks total, ", 0, " tasks running, ", 0, " tasks queued...");
check(pool.get_tasks_total() == 0 && pool.get_tasks_running() == 0 && pool.get_tasks_queued() == 0);
dual_println("Resetting pool to ", std::thread::hardware_concurrency(), " threads.");
pool.reset(std::thread::hardware_concurrency());
}
/**
* @brief Check that pausing works.
*/
void check_pausing()
{
BS::concurrency_t n = std::min<BS::concurrency_t>(std::thread::hardware_concurrency(), 4);
dual_println("Resetting pool to ", n, " threads.");
pool.reset(n);
dual_println("Pausing pool.");
pool.paused = true;
dual_println("Submitting ", n * 3, " tasks, each one waiting for 200ms.");
for (BS::concurrency_t i = 0; i < n * 3; ++i)
pool.push_task(
[i]
{
std::this_thread::sleep_for(std::chrono::milliseconds(200));
dual_println("Task ", i, " done.");
});
dual_println("Immediately after submission, should have: ", n * 3, " tasks total, ", 0, " tasks running, ", n * 3, " tasks queued...");
check(pool.get_tasks_total() == n * 3 && pool.get_tasks_running() == 0 && pool.get_tasks_queued() == n * 3);
std::this_thread::sleep_for(std::chrono::milliseconds(300));
dual_println("300ms later, should still have: ", n * 3, " tasks total, ", 0, " tasks running, ", n * 3, " tasks queued...");
check(pool.get_tasks_total() == n * 3 && pool.get_tasks_running() == 0 && pool.get_tasks_queued() == n * 3);
dual_println("Unpausing pool.");
pool.paused = false;
std::this_thread::sleep_for(std::chrono::milliseconds(300));
dual_println("300ms later, should have: ", n * 2, " tasks total, ", n, " tasks running, ", n, " tasks queued...");
check(pool.get_tasks_total() == n * 2 && pool.get_tasks_running() == n && pool.get_tasks_queued() == n);
dual_println("Pausing pool and using wait_for_tasks() to wait for the running tasks.");
pool.paused = true;
pool.wait_for_tasks();
dual_println("After waiting, should have: ", n, " tasks total, ", 0, " tasks running, ", n, " tasks queued...");
check(pool.get_tasks_total() == n && pool.get_tasks_running() == 0 && pool.get_tasks_queued() == n);
std::this_thread::sleep_for(std::chrono::milliseconds(200));
dual_println("200ms later, should still have: ", n, " tasks total, ", 0, " tasks running, ", n, " tasks queued...");
check(pool.get_tasks_total() == n && pool.get_tasks_running() == 0 && pool.get_tasks_queued() == n);
dual_println("Unpausing pool and using wait_for_tasks() to wait for all tasks.");
pool.paused = false;
pool.wait_for_tasks();
dual_println("After waiting, should have: ", 0, " tasks total, ", 0, " tasks running, ", 0, " tasks queued...");
check(pool.get_tasks_total() == 0 && pool.get_tasks_running() == 0 && pool.get_tasks_queued() == 0);
dual_println("Resetting pool to ", std::thread::hardware_concurrency(), " threads.");
pool.reset(std::thread::hardware_concurrency());
}
// ======================================
// Functions to verify exception handling
// ======================================
/**
* @brief Check that exception handling work.
*/
void check_exceptions()
{
bool caught = false;
std::future<void> my_future = pool.submit([] { throw std::runtime_error("Exception thrown!"); });
try
{
my_future.get();
}
catch (const std::exception& e)
{
if (e.what() == std::string("Exception thrown!"))
caught = true;
}
check(caught);
}
// =====================================
// Functions to verify vector operations
// =====================================
/**
* @brief Check that parallelized vector operations work as expected by calculating the sum of two randomized vectors of a specific size in two ways, single-threaded and multithreaded, and comparing the results.
*/
void check_vector_of_size(const size_t vector_size, const BS::concurrency_t num_tasks)
{
std::vector<int64_t> vector_1(vector_size);
std::vector<int64_t> vector_2(vector_size);
std::mt19937_64 mt(rd());
std::uniform_int_distribution<int64_t> vector_dist(-1000000, 1000000);
for (size_t i = 0; i < vector_size; ++i)
{
vector_1[i] = vector_dist(mt);
vector_2[i] = vector_dist(mt);
}
dual_println("Adding two vectors with ", vector_size, " elements using ", num_tasks, " tasks...");
std::vector<int64_t> sum_single(vector_size);
for (size_t i = 0; i < vector_size; ++i)
sum_single[i] = vector_1[i] + vector_2[i];
std::vector<int64_t> sum_multi(vector_size);
pool.parallelize_loop(
0, vector_size,
[&sum_multi, &vector_1, &vector_2](const size_t start, const size_t end)
{
for (size_t i = start; i < end; ++i)
sum_multi[i] = vector_1[i] + vector_2[i];
},
num_tasks)
.wait();
bool vectors_equal = true;
for (size_t i = 0; i < vector_size; ++i)
vectors_equal = vectors_equal && (sum_single[i] == sum_multi[i]);
check(vectors_equal);
}
/**
* @brief Check that parallelized vector operations work as expected by calculating the sum of two randomized vectors in two ways, single-threaded and multithreaded, and comparing the results.
*/
void check_vectors()
{
pool.reset();
std::mt19937_64 mt(rd());
std::uniform_int_distribution<size_t> size_dist(0, 1000000);
std::uniform_int_distribution<BS::concurrency_t> task_dist(1, pool.get_thread_count());
for (size_t i = 0; i < 10; ++i)
check_vector_of_size(size_dist(mt), task_dist(mt));
}
// ==================
// Main test function
// ==================
/**
* @brief Test that various aspects of the library are working as expected.
*/
void do_tests()
{
print_header("Checking that the constructor works:");
check_constructor();
print_header("Checking that reset() works:");
check_reset();
print_header("Checking that push_task() works:");
check_push_task();
print_header("Checking that submit() works:");
check_submit();
print_header("Checking that wait_for_tasks() works...");
check_wait_for_tasks();
print_header("Checking that parallelize_loop() works:");
check_parallelize_loop();
print_header("Checking that task monitoring works:");
check_task_monitoring();
print_header("Checking that pausing works:");
check_pausing();
print_header("Checking that exception handling works:");
check_exceptions();
print_header("Testing that vector operations produce the expected results:");
check_vectors();
}
// ==========================
// Functions for benchmarking
// ==========================
/**
* @brief Print the timing of a specific test.
*
* @param num_tasks The number of tasks.
* @param mean_sd An std::pair containing the mean as the first member and standard deviation as the second member.
*/
void print_timing(const BS::concurrency_t num_tasks, const std::pair<double, double>& mean_sd)
{
if (num_tasks == 0)
dual_print("Single-threaded");
else if (num_tasks == 1)
dual_print("With 1 task");
else
dual_print("With ", std::setw(4), num_tasks, " tasks");
dual_println(", mean execution time was ", std::setw(6), mean_sd.first, " ms with standard deviation ", std::setw(4), mean_sd.second, " ms.");
}
/**
* @brief Calculate and print the speedup obtained by multithreading.
*
* @param timings A vector of the timings corresponding to different numbers of tasks.
*/
void print_speedup(const std::vector<double>& timings, const BS::concurrency_t try_tasks[])
{
const std::vector<double>::const_iterator min_el = std::min_element(std::begin(timings), std::end(timings));
const double max_speedup = std::round(timings[0] / *min_el * 10) / 10;
const BS::concurrency_t num_tasks = try_tasks[min_el - std::begin(timings)];
dual_println("Maximum speedup obtained by multithreading vs. single-threading: ", max_speedup, "x, using ", num_tasks, " tasks.");
}
/**
* @brief Calculate the mean and standard deviation of a set of integers.
*
* @param timings The integers.
* @return An std::pair containing the mean as the first member and standard deviation as the second member.
*/
std::pair<double, double> analyze(const std::vector<std::chrono::milliseconds::rep>& timings)
{
double mean = 0;
for (size_t i = 0; i < timings.size(); ++i)
mean += static_cast<double>(timings[i]) / static_cast<double>(timings.size());
double variance = 0;
for (size_t i = 0; i < timings.size(); ++i)
variance += (static_cast<double>(timings[i]) - mean) * (static_cast<double>(timings[i]) - mean) / static_cast<double>(timings.size());
const double sd = std::sqrt(variance);
return std::pair(mean, sd);
}
/**
* @brief Generate a seed. The std::mt19937_64 in each task will be seeded using this function in order to avoid depleting the entropy of the random_device.
*
* @return A random unsigned 64-bit integer.
*/
uint64_t generate_seed()
{
static std::mt19937_64 mt(rd());
return mt();
}
/**
* @brief Benchmark multithreaded performance by generating random vectors.
*/
void check_performance()
{
// Reset the pool to ensure that we have a fresh start.
pool.reset();
// Set the formatting of floating point numbers.
dual_print(std::fixed, std::setprecision(1));
// Initialize a random distribution to randomize vectors with arbitrary floating point values.
const double range = std::sqrt(std::numeric_limits<double>::max());
std::uniform_real_distribution<double> vector_dist(-range, range);
// Initialize a timer object to measure execution time.
BS::timer tmr;
// Store the number of available hardware threads for easy access.
const BS::concurrency_t thread_count = pool.get_thread_count();
dual_println("Using ", thread_count, " threads.");
// Define the number of tasks to try in each run of the test (0 = single-threaded).
const BS::concurrency_t try_tasks[] = {0, thread_count / 4, thread_count / 2, thread_count, thread_count * 2, thread_count * 4};
// The size of the vectors to use for the test.
constexpr size_t vector_size = 500;
// How many times to repeat each run of the test in order to collect reliable statistics.
constexpr size_t repeat = 20;
dual_println("Each test will be repeated ", repeat, " times to collect reliable statistics.");
// The target duration of the single-threaded test in milliseconds. The total time spent on the test in the single-threaded case will be approximately equal to repeat * target_ms.
constexpr std::chrono::milliseconds::rep target_ms = 300;
// Vectors to store statistics.
std::vector<double> different_n_timings;
std::vector<std::chrono::milliseconds::rep> same_n_timings;
// Test how many vectors we need to generate to roughly achieve the target duration.
size_t num_vectors = 1;
do
{
num_vectors *= 2;
std::vector<std::vector<double>> vectors(num_vectors, std::vector<double>(vector_size));
std::mt19937_64 test_mt(rd());
tmr.start();
for (size_t i = 0; i < num_vectors; ++i)
{
for (size_t j = 0; j < vector_size; ++j)
vectors[i][j] = vector_dist(test_mt);
}
tmr.stop();
} while (tmr.ms() < target_ms);
num_vectors = static_cast<size_t>(std::llround(static_cast<double>(num_vectors) * static_cast<double>(target_ms) / static_cast<double>(tmr.ms())));
// Initialize the desired number of vectors.
std::vector<std::vector<double>> vectors(num_vectors, std::vector<double>(vector_size));
// Perform the test.
dual_println("\nGenerating ", num_vectors, " random vectors with ", vector_size, " elements each:");
for (BS::concurrency_t n : try_tasks)
{
for (size_t r = 0; r < repeat; ++r)
{
tmr.start();
if (n > 1)
{
pool.parallelize_loop(
0, num_vectors,
[&vector_dist, &vectors](const size_t start, const size_t end)
{
std::mt19937_64 multi_mt(generate_seed());
for (size_t i = start; i < end; ++i)
{
for (size_t j = 0; j < vector_size; ++j)
vectors[i][j] = vector_dist(multi_mt);
}
},
n)
.wait();
}
else
{
std::mt19937_64 single_mt(generate_seed());
for (size_t i = 0; i < num_vectors; ++i)
{
for (size_t j = 0; j < vector_size; ++j)
vectors[i][j] = vector_dist(single_mt);
}
}
tmr.stop();
same_n_timings.push_back(tmr.ms());
}
std::pair<double, double> mean_sd = analyze(same_n_timings);
print_timing(n, mean_sd);
different_n_timings.push_back(mean_sd.first);
same_n_timings.clear();
}
print_speedup(different_n_timings, try_tasks);
}
int main()
{
const std::string log_filename = "BS_thread_pool_test-" + get_time() + ".log";
if (output_log)
log_file.open(log_filename);
dual_println("A C++17 Thread Pool for High-Performance Scientific Computing");
dual_println("(c) 2022 Barak Shoshany (baraksh@gmail.com) (http://baraksh.com)");
dual_println("GitHub: https://github.com/bshoshany/thread-pool\n");
dual_println("Thread pool library version is ", BS_THREAD_POOL_VERSION, ".");
dual_println("Hardware concurrency is ", std::thread::hardware_concurrency(), ".");
if (output_log)
dual_println("Generating log file: ", log_filename, ".\n");
dual_println("Important: Please do not run any other applications, especially multithreaded applications, in parallel with this test!");
if (enable_tests)
do_tests();
if (tests_failed == 0)
{
if (enable_tests)
print_header("SUCCESS: Passed all " + std::to_string(tests_succeeded) + " checks!", '+');
if (enable_benchmarks)
{
print_header("Performing benchmarks:");
check_performance();
print_header("Thread pool performance test completed!", '+');
}
return EXIT_SUCCESS;
}
else
{
print_header("FAILURE: Passed " + std::to_string(tests_succeeded) + " checks, but failed " + std::to_string(tests_failed) + "!", '+');
dual_println("\nPlease submit a bug report at https://github.com/bshoshany/thread-pool/issues including the exact specifications of your system (OS, CPU, compiler, etc.) and the generated log file.");
return EXIT_FAILURE;
}
}
CHANGELOG.md
+136 -42
View File
@@ -1,56 +1,150 @@
# A C++17 Thread Pool for High-Performance Scientific Computing
[![DOI:10.5281/zenodo.4742687](https://zenodo.org/badge/DOI/10.5281/zenodo.4742687.svg)](https://doi.org/10.5281/zenodo.4742687)
[![arXiv:2105.00613](https://img.shields.io/badge/arXiv-2105.00613-b31b1b.svg)](https://arxiv.org/abs/2105.00613)
[![License: MIT](https://img.shields.io/github/license/bshoshany/thread-pool)](https://github.com/bshoshany/thread-pool/blob/master/LICENSE.txt)
![Language: C++17](https://img.shields.io/badge/Language-C%2B%2B17-yellow)
![File size in bytes](https://img.shields.io/github/size/bshoshany/thread-pool/BS_thread_pool.hpp)
![GitHub last commit](https://img.shields.io/github/last-commit/bshoshany/thread-pool)
[![GitHub repo stars](https://img.shields.io/github/stars/bshoshany/thread-pool?style=social)](https://github.com/bshoshany/thread-pool)
[![Twitter @BarakShoshany](https://img.shields.io/twitter/follow/BarakShoshany?style=social)](https://twitter.com/BarakShoshany)
[![Open in Visual Studio Code](https://img.shields.io/badge/-Open%20in%20Visual%20Studio%20Code-007acc)](https://vscode.dev/github/bshoshany/thread-pool)
# `BS::thread_pool`: a fast, lightweight, and easy-to-use C++17 thread pool library
By Barak Shoshany ([baraksh@gmail.com](mailto:baraksh@gmail.com)) ([https://baraksh.com/](https://baraksh.com/))
* [Version history](#version-history)
* [v3.0.0 (2022-05-30)](#v300-2022-05-30)
* [v2.0.0 (2021-08-14)](#v200-2021-08-14)
* [v1.9 (2021-07-29)](#v19-2021-07-29)
* [v1.8 (2021-07-28)](#v18-2021-07-28)
* [v1.7 (2021-06-02)](#v17-2021-06-02)
* [v1.6 (2021-05-26)](#v16-2021-05-26)
* [v1.5 (2021-05-07)](#v15-2021-05-07)
* [v1.4 (2021-05-05)](#v14-2021-05-05)
* [v1.3 (2021-05-03)](#v13-2021-05-03)
* [v1.2 (2021-04-29)](#v12-2021-04-29)
* [v1.1 (2021-04-24)](#v11-2021-04-24)
* [v1.0 (2021-01-15)](#v10-2021-01-15)
## Version history
* v2.0.0 (2021-08-14)
* From now on, version numbers will adhere to the [Semantic Versioning](https://semver.org/) specification in the format **major.minor.patch**.
* A file named `thread_pool_test.cpp` has been added to the package. It will perform automated tests of all aspects of the package, and benchmark some multithreaded matrix operations. Please run it on your system and [submit a bug report](https://github.com/bshoshany/thread-pool/issues) if any of the tests fail. In addition, the code is thoroughly documented, and is meant to serve as an extensive example of how to properly use the package.
* The package is now available through [vcpkg](https://github.com/microsoft/vcpkg). Instructions for how to install it have been added to `README.md`. See [this pull request](https://github.com/bshoshany/thread-pool/pull/18).
* The package now defines a macro `THREAD_POOL_VERSION`, which returns the version number and release date of the thread pool library as a string.
* `parallelize_loop()` has undergone some major changes (and is now incompatible with v1.x):
### v3.0.0 (2022-05-30)
* This is a major new release with many changes and improvements! Please note that code written using previous releases will need to be slightly modified to work with the new release. The changes needed to migrate to the new API are explicitly indicated below for your convenience.
* Breaking changes to the library header file:
* The header file has been renamed to `BS_thread_pool.hpp` to avoid potential conflict with other thread pool libraries.
* **API migration:** The library must now be included by invoking `#include "BS_thread_pool.hpp"`.
* All the definitions in the library, including the `thread_pool` class and the helper classes, are now located in the namespace `BS`. This namespace will also be used for my other C++ projects, and is intended to ensure consistency between my projects while avoiding potential name conflicts with other libraries.
* **API migration:** The thread pool class should now be invoked as `BS::thread_pool`. Alternatively, it is possible to employ `using BS::thread_pool` or even `using namespace BS` and then invoke `thread_pool` directly. Same for the `BS::synced_stream` and `BS::timer` helper classes.
* The macro `THREAD_POOL_VERSION`, which contains the version number and release date of the library, has been renamed to `BS_THREAD_POOL_VERSION` to avoid potential conflicts.
* **API migration:** The version must now be read from the macro `BS_THREAD_POOL_VERSION`.
* The public member `sleep_duration` has been removed. The thread pool now uses condition variables instead of sleep to facilitate waiting. This significantly improves performance (by 10%-50% in my testing), drastically decreases idle CPU utilization, and eliminates the need to set an optimal sleep time. This was a highly-requested change; see [issue #1](https://github.com/bshoshany/thread-pool/issues/1), [issue #12](https://github.com/bshoshany/thread-pool/issues/12), and [pull request #23](https://github.com/bshoshany/thread-pool/pull/23).
* **API migration:** Remove any code that relates to the public member `sleep_duration`.
* The template specializations for `submit()` have been merged. Now instead of two versions, one for functions with a return value and one for functions without a return value, there is just one version, which can accept any function. This makes the code more compact (and elegant). If a function with no return value is submitted, an `std::future<void>` is returned (the previous version returned an `std::future<bool>`)
* **API migration:** To wait for a task with no return value, simply call `wait()` or `get()` on the corresponding `std::future<void>`.
* `parallelize_loop()` now returns a future in the form of a new `BS::multi_future` helper class template. The member function `wait()` of this future allows waiting until all of the loop's blocks finish executing. In previous versions, calling `parallelize_loop()` both parallelized the loop and waited for the blocks to finish; now it is possible to do other stuff while the loop executes.
* **API migration:** Since `parallelize_loop()` no longer automatically blocks, you should either store the result in a `BS::multi_future` object and call its `wait()` member function, or simply call `parallelize_loop().wait()` to reproduce the old behavior.
* Non-breaking changes to the library header file:
* It is now possible to use `parallelize_loop()` with functions that have return values and get these values from all blocks at once through the `get()` member function of the `BS::multi_future`.
* The template specializations for `push_task()` have been merged. Now instead of two versions, one for functions with arguments and one for functions without arguments, there is just one version, which can accept any function.
* Constructors have been made `explicit`. See [issue #28](https://github.com/bshoshany/thread-pool/issues/28).
* `submit()` now uses `std::make_shared` instead of `new` to create the shared pointer. This means only one memory allocation is performed instead of two, which should improve performance. In addition, all unique pointers are now created using `std::make_unique`.
* A new helper class template, `BS::multi_future`, has been added. It's basically just a wrapper around `std::vector<std::future<T>>`. This class is used by the new implementation of `parallelize_loop()` to allow waiting for the entire loop, consisting of multiple tasks with their corresponding futures, to finish executing.
* `BS::multi_future` can also be used independently to handle multiple futures at once. For example, you can now keep track of several groups of tasks by storing their futures inside separate `BS::multi_future` objects and use either `wait()` to wait for all tasks in a specific group to finish or `get()` to get an `std::vector` with the return values of every task in the group.
* Integer types are now chosen in a smarter way to improve portability, allow for better compatibility with 32-bit systems, and prevent potential conversion errors.
* Added a new type, `BS::concurrency_t`, equal to the return type of `std::thread::hardware_concurrency()`. This is probably pointless, since the C++ standard requires this to be `unsigned int`, but it seems to me to make the code slightly more portable, in case some non-conforming compiler chooses to use a different integer type.
* C-style casts have been converted to C++ cast expressions for added clarity.
* Miscellaneous minor optimizations and style improvements.
* Changes to the test program:
* The program has been renamed to `BS_thread_pool_test.cpp` to avoid potential conflict with other thread pool libraries.
* The program now returns `EXIT_FAILURE` if any of the tests failed, for automation purposes. See [pull request #42](https://github.com/bshoshany/thread-pool/pull/42).
* Fixed incorrect check order in `check_task_monitoring()`. See [pull request #43](https://github.com/bshoshany/thread-pool/pull/43).
* Added a new test for `parallelize_loop()` with a return value.
* Improved some of the tests to make them more reliable. For example, `count_unique_threads()` now uses futures (stored in a `BS::multi_future<void>` object).
* The program now uses `std::vector` instead of matrices, for both consistency checks and benchmarks, in order to simplify the code and considerably reduce its length.
* The benchmarks have been simplified. There's now only one test: filling a specific number of vectors of fixed size with random values. This may be replaced with something more practical in a future released, but at least on the systems I've tested on, it does demonstrate a very significant multi-threading speedup.
* In addition to multi-threaded tests with different numbers of tasks, the benchmark now also includes a single-threaded test. This allows for more accurate benchmarks compared to previous versions, as the (slight) parallelization overhead is now taken into account when calculating the maximum speedup.
* The program decides how many vectors to use for benchmarking by testing how many are needed to reach a target duration in the single-threaded test. This ensures that the test takes approximately the same amount of time on different systems, and is thus more consistent and portable.
* Miscellaneous minor optimizations and style improvements.
* Changes to `README.md`:
* Many sections have been rewritten and/or polished.
* Explanations and examples of all the new features have been added.
* Added an acknowledgements section.
* Miscellaneous changes:
* Added a `CITATION.bib` file (in BibTeX format) to the GitHub repository. You can use it to easily cite this package if you use it in any research papers.
* Added a `CITATION.cff` file (in YAML format) to the GitHub repository. This should add [an option to get a citation in different formats](https://docs.github.com/en/github/creating-cloning-and-archiving-repositories/creating-a-repository-on-github/about-citation-files) directly from GitHub repository by clicking on "cite this repository" on the sidebar to the right.
* Added templates for GitHub issues and pull requests.
### v2.0.0 (2021-08-14)
* From now on, version numbers will adhere to the [Semantic Versioning](https://semver.org/) specification in the format **major.minor.patch**.
* A file named `thread_pool_test.cpp` has been added to the package. It will perform automated tests of all aspects of the package, and benchmark some multithreaded matrix operations. Please run it on your system and [submit a bug report](https://github.com/bshoshany/thread-pool/issues) if any of the tests fail. In addition, the code is thoroughly documented, and is meant to serve as an extensive example of how to properly use the package.
* The package is now available through [vcpkg](https://github.com/microsoft/vcpkg). Instructions for how to install it have been added to `README.md`. See [this pull request](https://github.com/bshoshany/thread-pool/pull/18).
* The package now defines a macro `THREAD_POOL_VERSION`, which returns the version number and release date of the thread pool library as a string.
* `parallelize_loop()` has undergone some major changes (and is now incompatible with v1.x):
* The second argument is now the index **after** the last index, instead of the last index itself. This is more consistent with C++ conventions (e.g. standard library algorithms) where the range is always `[first, last)`. For example, for an array with `n` indices, instead of `parallelize_loop(0, n - 1, ...)` you should now write `parallelize_loop(0, n, ...)`.
* The `loop` function is now only called once per block, instead of once per index, as was the case before. This should provide a performance boost due to significantly reducing the number of function calls, and it also allows you to conserve resources by using them only once per block instead of once per index (an example can be found in the `random_matrix_generator` class in `thread_pool_test.cpp`). It also means that `loop` now takes two arguments: the first index in the block and the index after the last index in the block. Thus, `loop(start, end)` should typically involve a loop of the form `for (T i = start; i < end; i++)`.
* The first and last indices can now be of two different integer types. Previously, `parallelize_loop(0, i, ...)` did not work if `i` was not an `int`, because `0` was interpreted as `int`, and the two arguments had to be of the same type. Therefore, one had to use casting, e.g. `parallelize_loop((size_t)0, i)`, to make it work. Now this is no longer necessary; the common type is inferred automatically using `std::common_type_t`.
* v1.9 (2021-07-29)
* Fixed a bug in `reset()` which caused it to create the wrong number of threads.
* v1.8 (2021-07-28)
* The version history has become too long to be included in `README.md`, so I moved it to a separate file, `CHANGELOG.md`.
* A button to open this repository directly in Visual Studio Code has been added to the badges in `README.md`.
* An internal variable named `promise` has been renamed to `task_promise` to avoid any potential errors in case the user invokes `using namespace std`.
* `submit()` now catches exceptions thrown by the submitted task and forwards them to the future. See [this issue](https://github.com/bshoshany/thread-pool/issues/14).
* Eliminated compiler warnings that appeared when using the `-Weffc++` flag in GCC. See [this pull request](https://github.com/bshoshany/thread-pool/pull/17).
* v1.7 (2021-06-02)
* Fixed a bug in `parallelize_loop()` which prevented it from actually running loops in parallel, see [this issue](https://github.com/bshoshany/thread-pool/issues/11).
* v1.6 (2021-05-26)
* Since MSVC does not interpret `and` as `&&` by default, the previous release did not compile with MSVC unless the `/permissive-` or `/Za` compiler flags were used. This has been fixed in this version, and the code now successfully compiles with GCC, Clang, and MSVC. See [this pull request](https://github.com/bshoshany/thread-pool/pull/10).
* v1.5 (2021-05-07)
* This library now has a DOI for citation purposes. Information on how to cite it in publications has been added to the source code and to `README.md`.
* Added GitHub badges to `README.md`.
* v1.4 (2021-05-05)
* Added three new public member functions to monitor the tasks submitted to the pool:
### v1.9 (2021-07-29)
* Fixed a bug in `reset()` which caused it to create the wrong number of threads.
### v1.8 (2021-07-28)
* The version history has become too long to be included in `README.md`, so I moved it to a separate file, `CHANGELOG.md`.
* A button to open this repository directly in Visual Studio Code has been added to the badges in `README.md`.
* An internal variable named `promise` has been renamed to `task_promise` to avoid any potential errors in case the user invokes `using namespace std`.
* `submit()` now catches exceptions thrown by the submitted task and forwards them to the future. See [this issue](https://github.com/bshoshany/thread-pool/issues/14).
* Eliminated compiler warnings that appeared when using the `-Weffc++` flag in GCC. See [this pull request](https://github.com/bshoshany/thread-pool/pull/17).
### v1.7 (2021-06-02)
* Fixed a bug in `parallelize_loop()` which prevented it from actually running loops in parallel, see [this issue](https://github.com/bshoshany/thread-pool/issues/11).
### v1.6 (2021-05-26)
* Since MSVC does not interpret `and` as `&&` by default, the previous release did not compile with MSVC unless the `/permissive-` or `/Za` compiler flags were used. This has been fixed in this version, and the code now successfully compiles with GCC, Clang, and MSVC. See [this pull request](https://github.com/bshoshany/thread-pool/pull/10).
### v1.5 (2021-05-07)
* This library now has a DOI for citation purposes. Information on how to cite it in publications has been added to the source code and to `README.md`.
* Added GitHub badges to `README.md`.
### v1.4 (2021-05-05)
* Added three new public member functions to monitor the tasks submitted to the pool:
* `get_tasks_queued()` gets the number of tasks currently waiting in the queue to be executed by the threads.
* `get_tasks_running()` gets the number of tasks currently being executed by the threads.
* `get_tasks_total()` gets the total number of unfinished tasks - either still in the queue, or running in a thread.
* Note that `get_tasks_running() == get_tasks_total() - get_tasks_queued()`.
* Renamed the private member variable `tasks_waiting` to `tasks_total` to make its purpose clearer.
* Added an option to temporarily pause the workers:
* Added an option to temporarily pause the workers:
* When public member variable `paused` is set to `true`, the workers temporarily stop popping new tasks out of the queue, although any tasks already executed will keep running until they are done. Set to `false` again to resume popping tasks.
* While the workers are paused, `wait_for_tasks()` will wait for the running tasks instead of all tasks (otherwise it would wait forever).
* By utilizing the new pausing mechanism, `reset()` can now change the number of threads on-the-fly while there are still tasks waiting in the queue. The new thread pool will resume executing tasks from the queue once it is created.
* `parallelize_loop()` and `wait_for_tasks()` now have the same behavior as the worker function with regards to waiting for tasks to complete. If the relevant tasks are not yet complete, then before checking again, they will sleep for `sleep_duration` microseconds, unless that variable is set to zero, in which case they will call `std::this_thread::yield()`. This should improve performance and reduce CPU usage.
* Merged [this commit](https://github.com/bshoshany/thread-pool/pull/8): Fixed weird error when using MSVC and including `windows.h`.
* The `README.md` file has been reorganized and expanded.
* v1.3 (2021-05-03)
* Fixed [this issue](https://github.com/bshoshany/thread-pool/issues/3): Removed `std::move` from the `return` statement in `push_task()`. This previously generated a `-Wpessimizing-move` warning in Clang. The assembly code generated by the compiler seems to be the same before and after this change, presumably because the compiler eliminates the `std::move` automatically, but this change gets rid of the Clang warning.
* Fixed [this issue](https://github.com/bshoshany/thread-pool/issues/5): Removed a debugging message printed to `std::cout`, which was left in the code by mistake.
* Fixed [this issue](https://github.com/bshoshany/thread-pool/issues/6): `parallelize_loop()` no longer sends references for the variables `start` and `stop` when calling `push_task()`, which may lead to undefined behavior.
* A companion paper is now published at <a href="https://arxiv.org/abs/2105.00613">arXiv:2105.00613</a>, including additional information such as performance tests on systems with up to 80 hardware threads. The `README.md` has been updated, and it is now roughly identical in content to the paper.
* v1.2 (2021-04-29)
* The worker function, which controls the execution of tasks by each thread, now sleeps by default instead of yielding. Previously, when the worker could not find any tasks in the queue, it called `std::this_thread::yield()` and then tried again. However, this caused the workers to have high CPU usage when idle, [as reported by some users](https://github.com/bshoshany/thread-pool/issues/1). Now, when the worker function cannot find a task to run, it instead sleeps for a duration given by the public member variable `sleep_duration` (in microseconds) before checking the queue again. The default value is `1000` microseconds, which I found to be optimal in terms of both CPU usage and performance, but your own optimal value may be different.
* If the constructor is called with an argument of zero for the number of threads, then the default value, `std::thread::hardware_concurrency()`, is used instead.
* Added a simple helper class, `timer`, which can be used to measure execution time for benchmarking purposes.
* Improved and expanded the documentation.
* v1.1 (2021-04-24)
* Cosmetic changes only. Fixed a typo in the Doxygen comments and added a link to the GitHub repository.
* v1.0 (2021-01-15)
* Initial release.
* `parallelize_loop()` and `wait_for_tasks()` now have the same behavior as the worker function with regards to waiting for tasks to complete. If the relevant tasks are not yet complete, then before checking again, they will sleep for `sleep_duration` microseconds, unless that variable is set to zero, in which case they will call `std::this_thread::yield()`. This should improve performance and reduce CPU usage.
* Merged [this commit](https://github.com/bshoshany/thread-pool/pull/8): Fixed weird error when using MSVC and including `windows.h`.
* The `README.md` file has been reorganized and expanded.
### v1.3 (2021-05-03)
* Fixed [this issue](https://github.com/bshoshany/thread-pool/issues/3): Removed `std::move` from the `return` statement in `push_task()`. This previously generated a `-Wpessimizing-move` warning in Clang. The assembly code generated by the compiler seems to be the same before and after this change, presumably because the compiler eliminates the `std::move` automatically, but this change gets rid of the Clang warning.
* Fixed [this issue](https://github.com/bshoshany/thread-pool/issues/5): Removed a debugging message printed to `std::cout`, which was left in the code by mistake.
* Fixed [this issue](https://github.com/bshoshany/thread-pool/issues/6): `parallelize_loop()` no longer sends references for the variables `start` and `stop` when calling `push_task()`, which may lead to undefined behavior.
* A companion paper is now published at <a href="https://arxiv.org/abs/2105.00613">arXiv:2105.00613</a>, including additional information such as performance tests on systems with up to 80 hardware threads. The `README.md` has been updated, and it is now roughly identical in content to the paper.
### v1.2 (2021-04-29)
* The worker function, which controls the execution of tasks by each thread, now sleeps by default instead of yielding. Previously, when the worker could not find any tasks in the queue, it called `std::this_thread::yield()` and then tried again. However, this caused the workers to have high CPU usage when idle, [as reported by some users](https://github.com/bshoshany/thread-pool/issues/1). Now, when the worker function cannot find a task to run, it instead sleeps for a duration given by the public member variable `sleep_duration` (in microseconds) before checking the queue again. The default value is `1000` microseconds, which I found to be optimal in terms of both CPU usage and performance, but your own optimal value may be different.
* If the constructor is called with an argument of zero for the number of threads, then the default value, `std::thread::hardware_concurrency()`, is used instead.
* Added a simple helper class, `timer`, which can be used to measure execution time for benchmarking purposes.
* Improved and expanded the documentation.
### v1.1 (2021-04-24)
* Cosmetic changes only. Fixed a typo in the Doxygen comments and added a link to the GitHub repository.
### v1.0 (2021-01-15)
* Initial release.
CITATION.bib
+13
View File
@@ -0,0 +1,13 @@
@article{Shoshany2021_ThreadPool,
archiveprefix = {arXiv},
author = {Barak Shoshany},
doi = {10.5281/zenodo.4742687},
eid = {arXiv:2105.00613},
eprint = {2105.00613},
journal = {arXiv e-prints},
keywords = {Computer Science - Distributed, Parallel, and Cluster Computing, D.1.3, D.1.5},
month = {May},
primaryclass = {cs.DC},
title = {{A C++17 Thread Pool for High-Performance Scientific Computing}},
year = {2021}
}
CITATION.cff
+24
View File
@@ -0,0 +1,24 @@
---
authors:
- family-names: "Shoshany"
given-names: "Barak"
orcid: "https://orcid.org/0000-0003-2222-127X"
cff-version: "1.2.0"
date-released: "2021-05-03"
doi: "10.5281/zenodo.4742687"
license: "MIT"
message: "If you use this package in published research, please cite it as follows."
repository-code: "https://github.com/bshoshany/thread-pool"
title: "A C++17 Thread Pool for High-Performance Scientific Computing"
preferred-citation:
type: "article"
authors:
- family-names: "Shoshany"
given-names: "Barak"
orcid: "https://orcid.org/0000-0003-2222-127X"
doi: "10.5281/zenodo.4742687"
journal: "arXiv"
month: 5
title: "A C++17 Thread Pool for High-Performance Scientific Computing"
url: "https://arxiv.org/abs/2105.00613"
year: 2021
LICENSE.txt
+1 -1
View File
@@ -1,6 +1,6 @@
MIT License
Copyright (c) 2021 Barak Shoshany
Copyright (c) 2022 Barak Shoshany
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
README.md
+401 -331
View File
File diff suppressed because it is too large Load Diff
thread_pool.hpp
-544
View File
@@ -1,544 +0,0 @@
#pragma once
/**
* @file thread_pool.hpp
* @author Barak Shoshany (baraksh@gmail.com) (http://baraksh.com)
* @version 2.0.0
* @date 2021-08-14
* @copyright Copyright (c) 2021 Barak Shoshany. Licensed under the MIT license. If you use this library in published research, please cite it as follows:
* - Barak Shoshany, "A C++17 Thread Pool for High-Performance Scientific Computing", doi:10.5281/zenodo.4742687, arXiv:2105.00613 (May 2021)
*
* @brief A C++17 thread pool for high-performance scientific computing.
* @details A modern C++17-compatible thread pool implementation, built from scratch with high-performance scientific computing in mind. The thread pool is implemented as a single lightweight and self-contained class, and does not have any dependencies other than the C++17 standard library, thus allowing a great degree of portability. In particular, this implementation does not utilize OpenMP or any other high-level multithreading APIs, and thus gives the programmer precise low-level control over the details of the parallelization, which permits more robust optimizations. The thread pool was extensively tested on both AMD and Intel CPUs with up to 40 cores and 80 threads. Other features include automatic generation of futures and easy parallelization of loops. Two helper classes enable synchronizing printing to an output stream by different threads and measuring execution time for benchmarking purposes. Please visit the GitHub repository at https://github.com/bshoshany/thread-pool for documentation and updates, or to submit feature requests and bug reports.
*/
#define THREAD_POOL_VERSION "v2.0.0 (2021-08-14)"
#include <atomic> // std::atomic
#include <chrono> // std::chrono
#include <cstdint> // std::int_fast64_t, std::uint_fast32_t
#include <functional> // std::function
#include <future> // std::future, std::promise
#include <iostream> // std::cout, std::ostream
#include <memory> // std::shared_ptr, std::unique_ptr
#include <mutex> // std::mutex, std::scoped_lock
#include <queue> // std::queue
#include <thread> // std::this_thread, std::thread
#include <type_traits> // std::common_type_t, std::decay_t, std::enable_if_t, std::is_void_v, std::invoke_result_t
#include <utility> // std::move
// ============================================================================================= //
// Begin class thread_pool //
/**
* @brief A C++17 thread pool class. The user submits tasks to be executed into a queue. Whenever a thread becomes available, it pops a task from the queue and executes it. Each task is automatically assigned a future, which can be used to wait for the task to finish executing and/or obtain its eventual return value.
*/
class thread_pool
{
typedef std::uint_fast32_t ui32;
typedef std::uint_fast64_t ui64;
public:
// ============================
// Constructors and destructors
// ============================
/**
* @brief Construct a new thread pool.
*
* @param _thread_count The number of threads to use. The default value is the total number of hardware threads available, as reported by the implementation. With a hyperthreaded CPU, this will be twice the number of CPU cores. If the argument is zero, the default value will be used instead.
*/
thread_pool(const ui32 &_thread_count = std::thread::hardware_concurrency())
: thread_count(_thread_count ? _thread_count : std::thread::hardware_concurrency()), threads(new std::thread[_thread_count ? _thread_count : std::thread::hardware_concurrency()])
{
create_threads();
}
/**
* @brief Destruct the thread pool. Waits for all tasks to complete, then destroys all threads. Note that if the variable paused is set to true, then any tasks still in the queue will never be executed.
*/
~thread_pool()
{
wait_for_tasks();
running = false;
destroy_threads();
}
// =======================
// Public member functions
// =======================
/**
* @brief Get the number of tasks currently waiting in the queue to be executed by the threads.
*
* @return The number of queued tasks.
*/
ui64 get_tasks_queued() const
{
const std::scoped_lock lock(queue_mutex);
return tasks.size();
}
/**
* @brief Get the number of tasks currently being executed by the threads.
*
* @return The number of running tasks.
*/
ui32 get_tasks_running() const
{
return tasks_total - (ui32)get_tasks_queued();
}
/**
* @brief Get the total number of unfinished tasks - either still in the queue, or running in a thread.
*
* @return The total number of tasks.
*/
ui32 get_tasks_total() const
{
return tasks_total;
}
/**
* @brief Get the number of threads in the pool.
*
* @return The number of threads.
*/
ui32 get_thread_count() const
{
return thread_count;
}
/**
* @brief Parallelize a loop by splitting it into blocks, submitting each block separately to the thread pool, and waiting for all blocks to finish executing. The user supplies a loop function, which will be called once per block and should iterate over the block's range.
*
* @tparam T1 The type of the first index in the loop. Should be a signed or unsigned integer.
* @tparam T2 The type of the index after the last index in the loop. Should be a signed or unsigned integer. If T1 is not the same as T2, a common type will be automatically inferred.
* @tparam F The type of the function to loop through.
* @param first_index The first index in the loop.
* @param index_after_last The index after the last index in the loop. The loop will iterate from first_index to (index_after_last - 1) inclusive. In other words, it will be equivalent to "for (T i = first_index; i < index_after_last; i++)". Note that if first_index == index_after_last, the function will terminate without doing anything.
* @param loop The function to loop through. Will be called once per block. Should take exactly two arguments: the first index in the block and the index after the last index in the block. loop(start, end) should typically involve a loop of the form "for (T i = start; i < end; i++)".
* @param num_blocks The maximum number of blocks to split the loop into. The default is to use the number of threads in the pool.
*/
template <typename T1, typename T2, typename F>
void parallelize_loop(const T1 &first_index, const T2 &index_after_last, const F &loop, ui32 num_blocks = 0)
{
typedef std::common_type_t<T1, T2> T;
T the_first_index = (T)first_index;
T last_index = (T)index_after_last;
if (the_first_index == last_index)
return;
if (last_index < the_first_index)
{
T temp = last_index;
last_index = the_first_index;
the_first_index = temp;
}
last_index--;
if (num_blocks == 0)
num_blocks = thread_count;
ui64 total_size = (ui64)(last_index - the_first_index + 1);
ui64 block_size = (ui64)(total_size / num_blocks);
if (block_size == 0)
{
block_size = 1;
num_blocks = (ui32)total_size > 1 ? (ui32)total_size : 1;
}
std::atomic<ui32> blocks_running = 0;
for (ui32 t = 0; t < num_blocks; t++)
{
T start = ((T)(t * block_size) + the_first_index);
T end = (t == num_blocks - 1) ? last_index + 1 : ((T)((t + 1) * block_size) + the_first_index);
blocks_running++;
push_task([start, end, &loop, &blocks_running]
{
loop(start, end);
blocks_running--;
});
}
while (blocks_running != 0)
{
sleep_or_yield();
}
}
/**
* @brief Push a function with no arguments or return value into the task queue.
*
* @tparam F The type of the function.
* @param task The function to push.
*/
template <typename F>
void push_task(const F &task)
{
tasks_total++;
{
const std::scoped_lock lock(queue_mutex);
tasks.push(std::function<void()>(task));
}
}
/**
* @brief Push a function with arguments, but no return value, into the task queue.
* @details The function is wrapped inside a lambda in order to hide the arguments, as the tasks in the queue must be of type std::function<void()>, so they cannot have any arguments or return value. If no arguments are provided, the other overload will be used, in order to avoid the (slight) overhead of using a lambda.
*
* @tparam F The type of the function.
* @tparam A The types of the arguments.
* @param task The function to push.
* @param args The arguments to pass to the function.
*/
template <typename F, typename... A>
void push_task(const F &task, const A &...args)
{
push_task([task, args...]
{ task(args...); });
}
/**
* @brief Reset the number of threads in the pool. Waits for all currently running tasks to be completed, then destroys all threads in the pool and creates a new thread pool with the new number of threads. Any tasks that were waiting in the queue before the pool was reset will then be executed by the new threads. If the pool was paused before resetting it, the new pool will be paused as well.
*
* @param _thread_count The number of threads to use. The default value is the total number of hardware threads available, as reported by the implementation. With a hyperthreaded CPU, this will be twice the number of CPU cores. If the argument is zero, the default value will be used instead.
*/
void reset(const ui32 &_thread_count = std::thread::hardware_concurrency())
{
bool was_paused = paused;
paused = true;
wait_for_tasks();
running = false;
destroy_threads();
thread_count = _thread_count ? _thread_count : std::thread::hardware_concurrency();
threads.reset(new std::thread[thread_count]);
paused = was_paused;
running = true;
create_threads();
}
/**
* @brief Submit a function with zero or more arguments and no return value into the task queue, and get an std::future<bool> that will be set to true upon completion of the task.
*
* @tparam F The type of the function.
* @tparam A The types of the zero or more arguments to pass to the function.
* @param task The function to submit.
* @param args The zero or more arguments to pass to the function.
* @return A future to be used later to check if the function has finished its execution.
*/
template <typename F, typename... A, typename = std::enable_if_t<std::is_void_v<std::invoke_result_t<std::decay_t<F>, std::decay_t<A>...>>>>
std::future<bool> submit(const F &task, const A &...args)
{
std::shared_ptr<std::promise<bool>> task_promise(new std::promise<bool>);
std::future<bool> future = task_promise->get_future();
push_task([task, args..., task_promise]
{
try
{
task(args...);
task_promise->set_value(true);
}
catch (...)
{
try
{
task_promise->set_exception(std::current_exception());
}
catch (...)
{
}
}
});
return future;
}
/**
* @brief Submit a function with zero or more arguments and a return value into the task queue, and get a future for its eventual returned value.
*
* @tparam F The type of the function.
* @tparam A The types of the zero or more arguments to pass to the function.
* @tparam R The return type of the function.
* @param task The function to submit.
* @param args The zero or more arguments to pass to the function.
* @return A future to be used later to obtain the function's returned value, waiting for it to finish its execution if needed.
*/
template <typename F, typename... A, typename R = std::invoke_result_t<std::decay_t<F>, std::decay_t<A>...>, typename = std::enable_if_t<!std::is_void_v<R>>>
std::future<R> submit(const F &task, const A &...args)
{
std::shared_ptr<std::promise<R>> task_promise(new std::promise<R>);
std::future<R> future = task_promise->get_future();
push_task([task, args..., task_promise]
{
try
{
task_promise->set_value(task(args...));
}
catch (...)
{
try
{
task_promise->set_exception(std::current_exception());
}
catch (...)
{
}
}
});
return future;
}
/**
* @brief Wait for tasks to be completed. Normally, this function waits for all tasks, both those that are currently running in the threads and those that are still waiting in the queue. However, if the variable paused is set to true, this function only waits for the currently running tasks (otherwise it would wait forever). To wait for a specific task, use submit() instead, and call the wait() member function of the generated future.
*/
void wait_for_tasks()
{
while (true)
{
if (!paused)
{
if (tasks_total == 0)
break;
}
else
{
if (get_tasks_running() == 0)
break;
}
sleep_or_yield();
}
}
// ===========
// Public data
// ===========
/**
* @brief An atomic variable indicating to the workers to pause. When set to true, the workers temporarily stop popping new tasks out of the queue, although any tasks already executed will keep running until they are done. Set to false again to resume popping tasks.
*/
std::atomic<bool> paused = false;
/**
* @brief The duration, in microseconds, that the worker function should sleep for when it cannot find any tasks in the queue. If set to 0, then instead of sleeping, the worker function will execute std::this_thread::yield() if there are no tasks in the queue. The default value is 1000.
*/
ui32 sleep_duration = 1000;
private:
// ========================
// Private member functions
// ========================
/**
* @brief Create the threads in the pool and assign a worker to each thread.
*/
void create_threads()
{
for (ui32 i = 0; i < thread_count; i++)
{
threads[i] = std::thread(&thread_pool::worker, this);
}
}
/**
* @brief Destroy the threads in the pool by joining them.
*/
void destroy_threads()
{
for (ui32 i = 0; i < thread_count; i++)
{
threads[i].join();
}
}
/**
* @brief Try to pop a new task out of the queue.
*
* @param task A reference to the task. Will be populated with a function if the queue is not empty.
* @return true if a task was found, false if the queue is empty.
*/
bool pop_task(std::function<void()> &task)
{
const std::scoped_lock lock(queue_mutex);
if (tasks.empty())
return false;
else
{
task = std::move(tasks.front());
tasks.pop();
return true;
}
}
/**
* @brief Sleep for sleep_duration microseconds. If that variable is set to zero, yield instead.
*
*/
void sleep_or_yield()
{
if (sleep_duration)
std::this_thread::sleep_for(std::chrono::microseconds(sleep_duration));
else
std::this_thread::yield();
}
/**
* @brief A worker function to be assigned to each thread in the pool. Continuously pops tasks out of the queue and executes them, as long as the atomic variable running is set to true.
*/
void worker()
{
while (running)
{
std::function<void()> task;
if (!paused && pop_task(task))
{
task();
tasks_total--;
}
else
{
sleep_or_yield();
}
}
}
// ============
// Private data
// ============
/**
* @brief A mutex to synchronize access to the task queue by different threads.
*/
mutable std::mutex queue_mutex = {};
/**
* @brief An atomic variable indicating to the workers to keep running. When set to false, the workers permanently stop working.
*/
std::atomic<bool> running = true;
/**
* @brief A queue of tasks to be executed by the threads.
*/
std::queue<std::function<void()>> tasks = {};
/**
* @brief The number of threads in the pool.
*/
ui32 thread_count;
/**
* @brief A smart pointer to manage the memory allocated for the threads.
*/
std::unique_ptr<std::thread[]> threads;
/**
* @brief An atomic variable to keep track of the total number of unfinished tasks - either still in the queue, or running in a thread.
*/
std::atomic<ui32> tasks_total = 0;
};
// End class thread_pool //
// ============================================================================================= //
// ============================================================================================= //
// Begin class synced_stream //
/**
* @brief A helper class to synchronize printing to an output stream by different threads.
*/
class synced_stream
{
public:
/**
* @brief Construct a new synced stream.
*
* @param _out_stream The output stream to print to. The default value is std::cout.
*/
synced_stream(std::ostream &_out_stream = std::cout)
: out_stream(_out_stream){};
/**
* @brief Print any number of items into the output stream. Ensures that no other threads print to this stream simultaneously, as long as they all exclusively use this synced_stream object to print.
*
* @tparam T The types of the items
* @param items The items to print.
*/
template <typename... T>
void print(const T &...items)
{
const std::scoped_lock lock(stream_mutex);
(out_stream << ... << items);
}
/**
* @brief Print any number of items into the output stream, followed by a newline character. Ensures that no other threads print to this stream simultaneously, as long as they all exclusively use this synced_stream object to print.
*
* @tparam T The types of the items
* @param items The items to print.
*/
template <typename... T>
void println(const T &...items)
{
print(items..., '\n');
}
private:
/**
* @brief A mutex to synchronize printing.
*/
mutable std::mutex stream_mutex = {};
/**
* @brief The output stream to print to.
*/
std::ostream &out_stream;
};
// End class synced_stream //
// ============================================================================================= //
// ============================================================================================= //
// Begin class timer //
/**
* @brief A helper class to measure execution time for benchmarking purposes.
*/
class timer
{
typedef std::int_fast64_t i64;
public:
/**
* @brief Start (or restart) measuring time.
*/
void start()
{
start_time = std::chrono::steady_clock::now();
}
/**
* @brief Stop measuring time and store the elapsed time since start().
*/
void stop()
{
elapsed_time = std::chrono::steady_clock::now() - start_time;
}
/**
* @brief Get the number of milliseconds that have elapsed between start() and stop().
*
* @return The number of milliseconds.
*/
i64 ms() const
{
return (std::chrono::duration_cast<std::chrono::milliseconds>(elapsed_time)).count();
}
private:
/**
* @brief The time point when measuring started.
*/
std::chrono::time_point<std::chrono::steady_clock> start_time = std::chrono::steady_clock::now();
/**
* @brief The duration that has elapsed between start() and stop().
*/
std::chrono::duration<double> elapsed_time = std::chrono::duration<double>::zero();
};
// End class timer //
// ============================================================================================= //
thread_pool_test.cpp
-1112
View File
File diff suppressed because it is too large Load Diff