thread pool
v4.1.0
BS::thread_pool : Une bibliothèque de piscine de thread rapide, légère et facile à utiliser Par Barak Shoshany
Courriel: [email protected]
Site Web: https://baraksh.com/
Github: https://github.com/bshoshany
Il s'agit de la documentation complète de la v4.1.0 de la bibliothèque, publiée le 2024-03-22.
BS::multi_future<T>BS::synced_streamBS::timerBS::signallerBS_thread_pool.hpp )BS::thread_poolBS::thread_poolBS::this_threadBS::multi_future<T>BS_thread_pool_utils.hpp )BS::signallerBS::synced_streamBS::timer Le multithreading est essentiel pour l'informatique moderne haute performance. Depuis C ++ 11, la bibliothèque standard C ++ a inclus la prise en charge de multithreading de bas niveau intégrée à l'aide de constructions telles que std::thread . Cependant, std::thread crée un nouveau thread à chaque fois qu'il est appelé, qui peut avoir une surcharge de performance significative. En outre, il est possible de créer plus de threads que le matériel ne peut gérer simultanément, ce qui entraîne un ralentissement substantiel.
La bibliothèque présentée ici contient une classe de pool de threads C ++, BS::thread_pool , qui évite ces problèmes en créant un pool fixe de threads une fois pour toutes, puis en réutilisant en continu les mêmes threads pour effectuer différentes tâches tout au long de la vie du programme. Par défaut, le nombre de threads dans le pool est égal au nombre maximum de threads que le matériel peut exécuter en parallèle.
L'utilisateur soumet les tâches à exécuter dans une file d'attente. Chaque fois qu'un fil devient disponible, il récupère la tâche suivante à partir de la file d'attente et l'exécute. Le pool produit automatiquement un std::future pour chaque tâche, ce qui permet à l'utilisateur d'attendre que la tâche termine l'exécution et / ou d'obtenir sa valeur de retour éventuelle, le cas échéant. Les threads et les tâches sont gérés de manière autonome par le pool en arrière-plan, sans nécessiter aucune entrée de l'utilisateur en plus de soumettre les tâches souhaitées.
La conception de cette bibliothèque est guidée par quatre principes importants. Tout d'abord, la compacité : la bibliothèque entière se compose d'un seul fichier d'en-tête autonome, sans autres composants ni dépendances, à part un petit fichier d'en-tête autonome avec des utilitaires facultatifs. Deuxièmement, la portabilité : la bibliothèque n'utilise que la bibliothèque standard C ++ 17, sans compter sur les extensions du compilateur ou les bibliothèques de 3e partie, et est donc compatible avec tout compilateur C ++ 17 moderne de conformité sur n'importe quelle plate-forme. Troisièmement, la facilité d'utilisation : la bibliothèque est largement documentée, et les programmeurs de tout niveau devraient pouvoir l'utiliser dès la sortie de la boîte.
Le quatrième et dernier principe directeur est les performances : chaque ligne de code de cette bibliothèque a été soigneusement conçue avec des performances maximales à l'esprit, et les performances ont été testées et vérifiées sur une variété de compilateurs et de plateformes. En effet, la bibliothèque a été initialement conçue pour être utilisée dans les propres projets informatiques scientifiques à forte intensité de calcul de l'auteur, exécutant à la fois sur des ordinateurs de bureau haut de gamme / ordinateur portable et des nœuds informatiques hautes performances.
D'autres bibliothèques multithreading plus avancées peuvent offrir plus de fonctionnalités et / ou de performances supérieures. Cependant, ils consistent généralement en une vaste base de code avec plusieurs composants et dépendances, et impliquent des API complexes qui nécessitent un investissement de temps substantiel pour apprendre. Cette bibliothèque n'est pas destinée à remplacer ces bibliothèques plus avancées; Au lieu de cela, il a été conçu pour les utilisateurs qui ne nécessitent pas de fonctionnalités très avancées, et préfèrent une bibliothèque simple et légère qui est facile à apprendre et à utiliser et qui peut être facilement intégrée dans les projets existants ou nouveaux.
#include "BS_thread_pool.hpp" et vous êtes prêt!submit_task() génère automatiquement un std::future , qui peut être utilisé pour attendre que la tâche termine l'exécution et / ou d'obtenir sa valeur de retour éventuelle.submit_loop() , qui renvoie un BS::multi_future qui peut être utilisé pour suivre l'exécution de toutes les tâches parallèles à la fois.detach_task() , et les boucles peuvent être parallélisées à l'aide de detach_loop() - Sacrifiant la commodité pour des performances encore plus importantes. Dans ce cas, wait() , wait_for() et wait_until() peuvent être utilisés pour attendre toutes les tâches de la file d'attente.BS_thread_pool_test.cpp peut être utilisé pour effectuer des tests et références automatisés exhaustifs, et sert également d'exemple complet de la façon d'utiliser correctement la bibliothèque. Le script PowerShell inclus BS_thread_pool_test.ps1 fournit un moyen portable d'exécuter les tests avec plusieurs compilateurs.BS_thread_pool_utils.hpp contient plusieurs classes d'utilité utiles.BS::signaller .BS::synced_stream .BS::timer .detach_sequence() et submit_sequence() .reset() .get_tasks_queued() , get_tasks_running() et des fonctions membres get_tasks_total() .get_thread_count() .pause() , unpause() et is_paused() ; Lorsqu'ils sont interrompus, les fils ne récupèrent pas de nouvelles tâches hors de la file d'attente.purge() .submit_task() ou submit_loop() à partir du fil principal via leur avenir.BS::this_thread::get_index() et un pointeur vers le pool qui possède le fil à l'aide de BS::this_thread::get_pool() .get_thread_ids() ou des poignées de thread définies par implémentation à l'aide de la fonction membre get_native_handles() en option.Cette bibliothèque doit compiler avec succès sur n'importe quel compilateur conforme à la norme C ++ 17, sur tous les systèmes d'exploitation et architectures pour lesquels un tel compilateur est disponible. La compatibilité a été vérifiée avec un processeur Intel I9-13900K à 24 cœurs (8p + 16e) / 32-thread à l'aide des compilateurs et plates-formes suivants:
De plus, cette bibliothèque a été testée sur un nœud Digital Research Alliance of Canada équipé de deux processeurs Intel Xeon Gold 6148 à 20 cœurs / 40-thread (pour un total de 40 cœurs et 80 threads), exécutant Centos Linux 7.9.2009, en utilisant GCC V13.2.0.
Le programme de test BS_thread_pool_test.cpp a été compilé sans avertissements (avec les drapeaux d'avertissement -Wall -Wextra -Wconversion -Wsign-conversion -Wpedantic -Weffc++ -Wshadow dans GCC / Clang et /W4 dans MSVC), exécuté et terminé tous les tests automatisés et les banclagés automatisés en utilisant tous les compilateurs et les systèmes ont réussi.
Comme cette bibliothèque nécessite des fonctionnalités C ++ 17, le code doit être compilé avec le support C ++ 17:
-std=c++17 . Sur Linux, vous devrez également utiliser l'indicateur -pthread pour activer la bibliothèque Posix Threads./std:c++17 , ainsi que /permissive- pour garantir la conformité des normes.Pour des performances maximales, il est recommandé de compiler avec toutes les optimisations du compilateur disponibles:
-O3 ./O2 . Par exemple, pour compiler le programme de test BS_thread_pool_test.cpp avec des avertissements et des optimisations, il est recommandé d'utiliser les commandes suivantes:
g++ BS_thread_pool_test.cpp -std=c++17 -O3 -Wall -Wextra -Wconversion -Wsign-conversion -Wpedantic -Weffc++ -Wshadow -pthread -o BS_thread_pool_testg++ par clang++ .-o BS_thread_pool_test par -o BS_thread_pool_test.exe et supprimer -pthread .cl BS_thread_pool_test.cpp /std:c++17 /permissive- /O2 /W4 /EHsc /Fo:BS_thread_pool_test.obj /Fe:BS_thread_pool_test.exe Pour installer BS::thread_pool , téléchargez simplement la dernière version du référentiel GitHub, placez le fichier d'en-tête BS_thread_pool.hpp à partir du dossier include dans le dossier souhaité et incluez-le dans votre programme:
# include " BS_thread_pool.hpp " Le pool de thread sera désormais accessible via la classe BS::thread_pool . Pour une installation encore plus rapide, vous pouvez télécharger le fichier d'en-tête lui-même directement sur cette URL.
Cette bibliothèque est également livrée avec un fichier d'en-tête indépendant des utilitaires BS_thread_pool_utils.hpp , qui n'est pas nécessaire pour utiliser le pool de threads, mais fournit des classes d'utilité qui peuvent être utiles pour le multithreading. Ce fichier d'en-tête réside également dans le dossier include . Il peut être téléchargé directement sur cette URL.
Cette bibliothèque est également disponible sur divers gestionnaires de packages et système de construction, notamment VCPKG, Conan, Meson et CMake avec CPM. Veuillez consulter ci-dessous pour plus de détails.
Le constructeur par défaut crée un pool de threads avec autant de threads que le matériel peut gérer simultanément, comme indiqué par l'implémentation via std::thread::hardware_concurrency() . Ceci est généralement déterminé par le nombre de noyaux dans le CPU. Si un noyau est hyperthread, il comptera comme deux threads. Par exemple:
// Constructs a thread pool with as many threads as available in the hardware.
BS::thread_pool pool;Facultativement, un certain nombre de threads différents de la concurrence matérielle peuvent être spécifiés comme argument au constructeur. Cependant, notez que l'ajout de threads que le matériel ne peut gérer ne s'améliorera pas , et en fait, le gênera très probablement. Cette option existe afin de permettre d'utiliser moins de threads que la concurrence matérielle, dans les cas où vous souhaitez laisser certains threads disponibles pour d'autres processus. Par exemple:
// Constructs a thread pool with only 12 threads.
BS::thread_pool pool ( 12 );Habituellement, lorsque le pool de threads est utilisé, le thread principal d'un programme ne doit soumettre que des tâches au pool de threads et attendre qu'ils se terminent, et ne doivent effectuer aucune tâche intensive en calcul en soi. Dans ce cas, il est recommandé d'utiliser la valeur par défaut pour le nombre de threads. Cela garantit que tous les fils disponibles dans le matériel seront mis au travail pendant que le fil principal attend.
La fonction membre get_thread_count() renvoie le nombre de threads dans le pool. Cela sera égal à std::thread::hardware_concurrency() si le constructeur par défaut a été utilisé.
Il n'est généralement pas nécessaire de modifier le nombre de threads dans la piscine après sa création, car tout le point d'un pool de threads est que vous ne créez les threads qu'une seule fois. Cependant, si nécessaire, cela peut être fait, en toute sécurité et à la volée, en utilisant la fonction membre reset() .
reset() attendra que toutes les tâches en cours d'exécution seront terminées, mais laisseront le reste des tâches dans la file d'attente. Ensuite, il détruira le pool de threads et en créera un nouveau avec le nouveau nombre de threads souhaité, comme spécifié dans l'argument de la fonction (ou la concurrence matérielle si aucun argument n'est donné). Le nouveau pool de threads reprendra ensuite l'exécution des tâches qui sont restées dans la file d'attente et de toute nouvelle tâche soumise.
Si vous le souhaitez, la version de cette bibliothèque peut être lue pendant le temps de compilation à partir des trois macros suivantes:
BS_THREAD_POOL_VERSION_MAJOR - indique la version principale.BS_THREAD_POOL_VERSION_MINOR - indique la version mineure.BS_THREAD_POOL_VERSION_PATCH - indique la version du patch. std::cout << " Thread pool library version is " << BS_THREAD_POOL_VERSION_MAJOR << ' . ' << BS_THREAD_POOL_VERSION_MINOR << ' . ' << BS_THREAD_POOL_VERSION_PATCH << " . n " ;Exemple de sortie:
Thread pool library version is 4.1.0.
Cela peut être utilisé, par exemple, pour permettre à la même base de code de fonctionner avec plusieurs versions incompatibles de la bibliothèque à l'aide des directives #if .
Remarque: cette fonctionnalité n'est disponible qu'à partir de v4.0.1. Les versions antérieures de cette bibliothèque ne définissent pas ces macros.
Dans cette section, nous apprendrons à soumettre une tâche sans arguments, mais potentiellement avec une valeur de retour, dans la file d'attente. Une fois qu'une tâche a été soumise, elle sera exécutée dès qu'un thread sera disponible. Les tâches sont exécutées dans l'ordre où ils ont été soumis (premier-in, premier-out), sauf si la priorité de la tâche est activée (voir ci-dessous).
Par exemple, si le pool a 8 threads et une file d'attente vide, et que nous avons soumis 16 tâches, nous devons nous attendre à ce que les 8 premières tâches soient exécutées en parallèle, les tâches restantes étant ramassées par les fils un par un alors que chaque fil termine sa première tâche, jusqu'à ce qu'aucune tâche ne soit laissée dans la file d'attente.
La fonction membre submit_task() est utilisée pour soumettre des tâches à la file d'attente. Il prend exactement une entrée, la tâche à soumettre. Cette tâche doit être une fonction sans arguments, mais elle peut avoir une valeur de retour. La valeur de retour est un std::future associé à la tâche.
Si la fonction soumise a une valeur de retour de type T , l'avenir sera de type std::future<T> , et sera défini sur la valeur de retour lorsque la fonction terminera son exécution. Si la fonction soumise n'a pas de valeur de retour, l'avenir sera un std::future<void> , qui ne renverra aucune valeur mais pourrait toujours être utilisé pour attendre la fin de la fonction.
L'utilisation de auto pour la valeur de retour de submit_task() signifie que le compilateur détectera automatiquement quelle instance du modèle std::future à utiliser. Cependant, la spécification du type particulier std::future<T> , comme dans les exemples ci-dessous, est recommandée pour une lisibilité accrue.
Pour attendre la fin de la tâche, utilisez la fonction membre wait() du futur. Pour obtenir la valeur de retour, utilisez la fonction membre get() , qui attendra également automatiquement que la tâche se termine si elle ne l'a pas encore été. Voici un exemple simple:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < future > // std::future
# include < iostream > // std::cout
int the_answer ()
{
return 42 ;
}
int main ()
{
BS::thread_pool pool;
std::future< int > my_future = pool. submit_task (the_answer);
std::cout << my_future. get () << ' n ' ;
} Dans cet exemple, nous avons soumis la fonction the_answer() , qui renvoie un int . La fonction membre submit_task() du pool a donc renvoyé un std::future<int> . Nous avons ensuite utilisé Utilisé la fonction membre get() du futur pour obtenir la valeur de retour et l'avons imprimée.
En plus de soumettre une fonction prédéfinie, nous pouvons également utiliser une expression de lambda pour définir rapidement la tâche à la volée. Réécriture de l'exemple précédent en termes d'expression de lambda, nous obtenons:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < future > // std::future
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool;
std::future< int > my_future = pool. submit_task ([]{ return 42 ; });
std::cout << my_future. get () << ' n ' ;
} Ici, l'expression lambda []{ return 42; } a deux parties:
[] . Cela signifie pour le compilateur qu'une expression de lambda est définie.{ return 42; } qui renvoie simplement la valeur 42 .Il est généralement plus simple et plus rapide de soumettre des expressions lambda plutôt que des fonctions prédéfinies, en particulier en raison de la capacité de capturer des variables locales, dont nous discuterons dans la section suivante.
Bien sûr, les tâches n'ont pas à renvoyer des valeurs. Dans l'exemple suivant, nous soumettons une fonction sans valeur de retour, puis utilisons l'avenir pour attendre qu'il termine l'exécution:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < chrono > // std::chrono
# include < future > // std::future
# include < iostream > // std::cout
# include < thread > // std::this_thread
int main ()
{
BS::thread_pool pool;
const std::future< void > my_future = pool. submit_task (
[]
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 500 ));
});
std::cout << " Waiting for the task to complete... " ;
my_future. wait ();
std::cout << " Done. " << ' n ' ;
} Ici, nous divisons la Lambda en plusieurs lignes pour la rendre plus lisible. La commande std::this_thread::sleep_for(std::chrono::milliseconds(500)) demande à la tâche de simplement dormir pendant 500 millisecondes, simulant une tâche à forte intensité de calcul.
Comme indiqué dans la section précédente, les tâches soumises à l'aide de submit_task() ne peuvent avoir aucun argument. Cependant, il est facile de soumettre des tâches avec un argument soit en emballage la fonction dans un lambda ou en utilisant directement les captures de Lambda. Voici deux exemples.
Ce qui suit est un exemple de soumission d'une fonction prédéfinie avec des arguments en l'emballant avec une lambda:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < future > // std::future
# include < iostream > // std::cout
double multiply ( const double lhs, const double rhs)
{
return lhs * rhs;
}
int main ()
{
BS::thread_pool pool;
std::future< double > my_future = pool. submit_task (
[]
{
return multiply ( 6 , 7 );
});
std::cout << my_future. get () << ' n ' ;
} Comme vous pouvez le voir, pour passer les arguments à multiply nous avons simplement appelé multiply(6, 7) explicitement à l'intérieur d'une lambda. Si les arguments ne sont pas des littéraux, nous devons utiliser la clause de capture lambda pour capturer les arguments de la portée locale:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < future > // std::future
# include < iostream > // std::cout
double multiply ( const double lhs, const double rhs)
{
return lhs * rhs;
}
int main ()
{
BS::thread_pool pool;
constexpr double first = 6 ;
constexpr double second = 7 ;
std::future< double > my_future = pool. submit_task (
[first, second]
{
return multiply (first, second);
});
std::cout << my_future. get () << ' n ' ;
} Nous pourrions même nous débarrasser entièrement de la fonction multiply et tout mettre dans une lambda, si vous le souhaitez:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < future > // std::future
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool;
constexpr double first = 6 ;
constexpr double second = 7 ;
std::future< double > my_future = pool. submit_task (
[first, second]
{
return first * second;
});
std::cout << my_future. get () << ' n ' ;
} Habituellement, il est préférable de soumettre une tâche à la file d'attente à l'aide de submit_task() . Cela vous permet d'attendre que la tâche termine et / ou d'obtenir sa valeur de retour plus tard. Cependant, parfois un avenir n'est pas nécessaire, par exemple lorsque vous souhaitez simplement "définir et oublier" une certaine tâche, ou si la tâche communique déjà avec le thread principal ou avec d'autres tâches sans utiliser de futures, comme les variables de condition.
Dans de tels cas, vous souhaiterez peut-être éviter les frais généraux impliqués dans l'attribution d'un avenir à la tâche afin d'augmenter les performances. C'est ce qu'on appelle «détacher» la tâche, car la tâche se détache du fil principal et s'exécute indépendamment.
Les tâches de détachement se font à l'aide de la fonction membre detach_task() , qui vous permet de détacher une tâche à la file d'attente sans générer un avenir pour cela. La tâche peut avoir n'importe quel nombre d'arguments, mais elle ne peut pas avoir de valeur de retour, car il n'y aurait aucun moyen pour le thread principal de récupérer cette valeur.
Étant donné que detach_task() ne renvoie pas d'avenir, il n'y a aucun moyen intégré pour l'utilisateur de savoir quand la tâche termine l'exécution. Vous devez vous assurer manuellement que la tâche termine l'exécution avant d'essayer d'utiliser tout ce qui dépend de sa sortie. Sinon, de mauvaises choses se produiront!
BS::thread_pool fournit à la fonction du membre wait() pour faciliter l'attente de toutes les tâches de la file d'attente, qu'elles soient détachées ou soumises avec un avenir. La fonction de membre wait() fonctionne de manière similaire à la fonction de membre wait() de std::future . Considérez, par exemple, le code suivant:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < chrono > // std::chrono
# include < iostream > // std::cout
# include < thread > // std::this_thread
int main ()
{
BS::thread_pool pool;
int result = 0 ;
pool. detach_task (
[&result]
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 100 ));
result = 42 ;
});
std::cout << result << ' n ' ;
} Ce programme définit d'abord une variable locale nommée result et l'initialise à 0 . Il détache ensuite une tâche sous la forme d'une expression de lambda. Notez que le Lambda capture result par référence , comme indiqué par le & devant. Cela signifie que la tâche peut modifier result et toute modification de cette telle sera reflétée dans le thread principal. Les changements de tâche result par 42 , mais il dort d'abord pendant 100 millisecondes. Lorsque le thread principal imprime la valeur du result , la tâche n'a pas encore eu le temps de modifier sa valeur, car elle dort toujours. Par conséquent, le programme imprimera la valeur initiale 0 .
Pour attendre que la tâche se termine, nous devons utiliser la fonction de membre wait() après l'avoir détachée:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < chrono > // std::chrono
# include < iostream > // std::cout
# include < thread > // std::this_thread
int main ()
{
BS::thread_pool pool;
int result = 0 ;
pool. detach_task (
[&result]
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 100 ));
result = 42 ;
});
pool. wait ();
std::cout << result << ' n ' ;
} Maintenant, le programme imprimera la valeur 42 , comme prévu. Notez cependant que wait() attendra toutes les tâches de la file d'attente, y compris toutes les autres tâches qui ont été potentiellement soumises avant ou après celle qui nous tient à cœur. Si nous voulons attendre une seule tâche, submit_task() serait un meilleur choix.
Parfois, vous pouvez attendre que les tâches se terminent, mais seulement pour un certain temps, ou jusqu'à un moment spécifique. Par exemple, si les tâches ne se sont pas encore terminées après un certain temps, vous souhaiterez peut-être faire savoir à l'utilisateur qu'il y a un retard.
Pour la tâche soumise avec des contrats à terme à l'aide submit_task() , cela peut être réalisé en utilisant deux fonctions membres de std::future :
wait_for() attend que la tâche soit terminée, mais cesse d'attendre après la durée spécifiée, donnée comme argument de type std::chrono::duration , est passée.wait_until() attend que la tâche soit terminée, mais arrête d'attendre après le moment spécifié, donné comme argument de type std::chrono::time_point , a été atteint. Dans les deux cas, les fonctions renverront future_status::ready si l'avenir est prêt, ce qui signifie que la tâche est terminée et sa valeur de retour, le cas échéant, a été obtenue. Cependant, il renverra std::future_status::timeout si l'avenir n'est pas encore prêt au moment où le délai d'expiration a expiré.
Voici un exemple:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < chrono > // std::chrono
# include < future > // std::future
# include < iostream > // std::cout
# include < thread > // std::this_thread
int main ()
{
BS::thread_pool pool;
const std::future< void > my_future = pool. submit_task (
[]
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
std::cout << " Task done! n " ;
});
while ( true )
{
if (my_future. wait_for ( std::chrono::milliseconds ( 200 )) != std::future_status::ready)
std::cout << " Sorry, the task is not done yet. n " ;
else
break ;
}
}La sortie doit ressembler à ceci:
Sorry, the task is not done yet.
Sorry, the task is not done yet.
Sorry, the task is not done yet.
Sorry, the task is not done yet.
Task done!
Pour les tâches détachées, car nous n'avons pas d'avenir pour eux, nous ne pouvons pas utiliser cette méthode. Cependant, BS::thread_pool a deux fonctions membres, également nommé wait_for() et wait_until() , qui attendent de la même manière une durée spécifiée ou jusqu'à un moment spécifié, mais faites-le pour toutes les tâches (soumises ou détachées). Au lieu d'un std::future_status , les fonctions d'attente du pool de threads renvoient true si toutes les tâches finies en cours d'exécution, ou false si la durée a expiré ou que le point dans le temps a été atteint mais que certaines tâches sont toujours en cours d'exécution.
Voici le même exemple que ci-dessus, en utilisant detach_task() et pool.wait_for() :
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < chrono > // std::chrono
# include < iostream > // std::cout
# include < thread > // std::this_thread
int main ()
{
BS::thread_pool pool;
pool. detach_task (
[]
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
std::cout << " Task done! n " ;
});
while ( true )
{
if (!pool. wait_for ( std::chrono::milliseconds ( 200 )))
std::cout << " Sorry, the task is not done yet. n " ;
else
break ;
}
}Examinons le programme suivant:
# include < iostream > // std::cout, std::boolalpha
class flag_class
{
public:
[[nodiscard]] bool get_flag () const
{
return flag;
}
void set_flag ( const bool arg)
{
flag = arg;
}
private:
bool flag = false ;
};
int main ()
{
flag_class flag_object;
flag_object. set_flag ( true );
std::cout << std::boolalpha << flag_object. get_flag () << ' n ' ;
} Ce programme crée un nouvel objet flag_object de la classe flag_class , définit l'indicateur sur true en utilisant la fonction membre du secteur set_flag() , puis imprime la valeur de l'indicateur à l'aide de la fonction de membre Getter get_flag() .
Et si nous voulons soumettre la fonction membre set_flag() comme tâche au pool de threads? Nous enroulons simplement l'instruction entière flag_object.set_flag(true); De la ligne dans un lambda, et passez flag_object à la Lambda par référence, comme dans cet exemple:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iostream > // std::cout, std::boolalpha
class flag_class
{
public:
[[nodiscard]] bool get_flag () const
{
return flag;
}
void set_flag ( const bool arg)
{
flag = arg;
}
private:
bool flag = false ;
};
int main ()
{
BS::thread_pool pool;
flag_class flag_object;
pool. submit_task (
[&flag_object]
{
flag_object. set_flag ( true );
})
. wait ();
std::cout << std::boolalpha << flag_object. get_flag () << ' n ' ;
} Bien sûr, cela fonctionnera également avec detach_task() , si nous appelons wait() sur la piscine elle-même au lieu de l'avenir retourné.
Notez que dans cet exemple, au lieu d'obtenir un avenir à partir de submit_task() , puis d'attendre cet avenir, nous avons simplement appelé wait() sur cet avenir immédiatement. C'est un moyen courant d'attendre une tâche à terminer si nous n'avons rien d'autre à faire entre-temps. Notez également que nous avons passé flag_object par référence au lambda, car nous voulons définir l'indicateur sur ce même objet, pas une copie de celui-ci (passer par valeur n'aurait pas fonctionné de toute façon, car les variables capturées par valeur sont implicitement const ).
Une autre chose que vous voudrez peut-être faire est d'appeler une fonction membre à partir de l'objet lui-même, c'est-à-dire à partir d'une autre fonction membre. Cela suit une syntaxe similaire, sauf que vous devez également capturer this (c'est-à-dire un pointeur vers l'objet actuel) dans la lambda. Voici un exemple:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iostream > // std::cout, std::boolalpha
BS::thread_pool pool;
class flag_class
{
public:
[[nodiscard]] bool get_flag () const
{
return flag;
}
void set_flag ( const bool arg)
{
flag = arg;
}
void set_flag_to_true ()
{
pool. submit_task (
[ this ]
{
set_flag ( true );
})
. wait ();
}
private:
bool flag = false ;
};
int main ()
{
flag_class flag_object;
flag_object. set_flag_to_true ();
std::cout << std::boolalpha << flag_object. get_flag () << ' n ' ;
} Notez que dans cet exemple, nous avons défini le pool de threads comme un objet global, afin qu'il soit accessible en dehors de la fonction main() .
L'une des méthodes de parallélisation les plus courantes et les plus efficaces est de diviser une boucle en boucles plus petites et de les exécuter en parallèle. Il est plus efficace dans les calculs "embarrassants parallèles", tels que les opérations vectorielles ou matricielles, où chaque itération de la boucle est complètement indépendante de toutes les autres itations.
Par exemple, si nous résumons deux vecteurs de 1000 éléments chacun, et que nous avons 10 threads, nous pourrions diviser la sommation en 10 blocs de 100 éléments chacun, et exécuter tous les blocs en parallèle et potentiellement augmenter les performances jusqu'à un facteur de 10.
BS::thread_pool peut automatiquement paralléliser les boucles. Pour voir comment cela fonctionne, considérez la boucle générique suivante:
for (T i = start; i < end; ++i)
loop (i);où:
T est tout type entier signé ou non signé.[start, end) , c'est-à-dire y compris start mais à l'exclusion de end .loop() est une opération effectuée pour chaque index de boucle i , tel que la modification d'un tableau avec des éléments end - start . Cette boucle peut être automatiquement parallélisée et soumise à la file d'attente du pool de threads à l'aide de la fonction membre submit_loop() , qui a la syntaxe suivante:
pool.submit_loop(start, end, loop, num_blocks);où:
start est le premier index de la plage.end est l'index après le dernier index de la plage, de sorte que la plage complète est [start, end) . En d'autres termes, la boucle sera équivalente à celle ci-dessus si start et end sont les mêmes.start et end doivent être tous deux du même type entier T Voir ci-dessous pour des exemples de ce qu'il faut faire lorsqu'ils ne sont pas du même type.end <= start , rien ne se passera.loop() est la fonction qui devrait s'exécuter dans chaque itération de la boucle et prend un argument, l'index de boucle.num_blocks est le nombre de blocs de la forme [a, b) pour diviser la boucle en. Par exemple, si la plage est [0, 9) et qu'il y a 3 blocs, les blocs seront les plages [0, 3) , [3, 6) et [6, 9) .[0, 100) est divisée en 15 blocs, le résultat sera de 10 blocs de taille 7, qui sera exécuté en premier et 5 blocs de taille 6.Chaque bloc sera soumis à la file d'attente du pool de threads en tant que tâche distincte. Par conséquent, une boucle divisée en 3 blocs sera divisée en 3 tâches individuelles, qui peuvent fonctionner en parallèle. S'il n'y a qu'un seul bloc, la boucle entière s'exécutera en une seule tâche et aucune parallélisation n'aura lieu.
Pour paralléliser la boucle générique ci-dessus, nous utilisons les commandes suivantes:
BS::multi_future< void > loop_future = pool.submit_loop(start, end, loop, num_blocks);
loop_future.wait(); submit_loop() Renvoie un objet du modèle de classe d'assistance BS::multi_future . Il s'agit essentiellement d'une spécialisation de std::vector<std::future<T>> avec des fonctions de membres supplémentaires. Chacun des blocs num_blocks aura un std::future lui sera attribué, et tous ces futurs seront stockés à l'intérieur du BS::multi_future retourné. Lorsque loop_future.wait() est appelé, le thread principal attendra que toutes les tâches générées par submit_loop() finissent l'exécution, et seules ces tâches - pas d'autres tâches qui se trouvent également dans la file d'attente. Il s'agit essentiellement du rôle de la classe BS::multi_future : attendre un groupe spécifique de tâches , dans ce cas, les tâches exécutant les blocs de boucle.
Quelle valeur devez-vous utiliser pour num_blocks ? L'omission de cet argument, de sorte que le nombre de blocs sera égal au nombre de threads dans la piscine, est généralement un bon choix. Pour les meilleures performances, il est recommandé de faire vos propres repères pour trouver le nombre optimal de blocs pour chaque boucle (vous pouvez utiliser la classe d'utilité BS::timer ). L'utilisation de moins de tâches qu'il y a de threads peut être préférée si vous exécutez également d'autres tâches en parallèle. L'utilisation de plus de tâches que de threads peut améliorer les performances dans certains cas, mais la parallélisation avec trop de tâches souffrira de rendements décroissants.
À titre d'exemple simple, le code suivant calcule et imprime un tableau de carrés de tous les entiers de 0 à 99:
# include < iomanip > // std::setw
# include < iostream > // std::cout
int main ()
{
constexpr unsigned int max = 100 ;
unsigned int squares[max];
for ( unsigned int i = 0 ; i < max; ++i)
squares[i] = i * i;
for ( unsigned int i = 0 ; i < max; ++i)
std::cout << std::setw ( 2 ) << i << " ^2 = " << std::setw ( 4 ) << squares[i] << ((i % 5 != 4 ) ? " | " : " n " );
}Nous pouvons le paralléliser comme suit:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iomanip > // std::setw
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool ( 10 );
constexpr unsigned int max = 100 ;
unsigned int squares[max];
const BS::multi_future< void > loop_future = pool. submit_loop < unsigned int >( 0 , max,
[&squares]( const unsigned int i)
{
squares[i] = i * i;
});
loop_future. wait ();
for ( unsigned int i = 0 ; i < max; ++i)
std::cout << std::setw ( 2 ) << i << " ^2 = " << std::setw ( 4 ) << squares[i] << ((i % 5 != 4 ) ? " | " : " n " );
} Puisqu'il y a 10 threads et que nous avons omis l'argument num_blocks , la boucle sera divisée en 10 blocs, chacun calculant 10 carrés.
Notez que submit_loop() a été exécuté avec le paramètre de modèle explicite <unsigned int> . La raison en est que les deux indices de boucle doivent être du même type. Cependant, ici, max est un unsigned int , tandis que 0 est un int (signé), donc les types ne correspondent pas, et le code ne se compilera que si nous obligeons le 0 à être du bon type. Cela peut être fait le plus élégamment en spécifiant le type d'indices explicitement à l'aide du paramètre de modèle.
La raison pour laquelle cela ne se fait pas automatiquement (par exemple, en utilisant std::common_type , c'est qu'il peut entraîner des indices négatifs accidentels à un type non signé, ou des indices entiers à un type entier trop étroit, ce qui peut conduire à une plage de boucle incorrecte.
Nous pourrions également lancer le 0 explicitement à unsigned int, mais cela n'a pas l'air aussi agréable:
pool.submit_loop( static_cast < unsigned int >( 0 ), max, /* ... */ );Ou nous pourrions utiliser un casting de style C:
pool.submit_loop(( unsigned int )( 0 ), max, /* ... */ );Ou nous pourrions utiliser un suffixe littéral entier:
pool.submit_loop< size_t >( 0U , max, ...);Dans une note secondaire, notez que nous parallélins ici le calcul des carrés, mais nous n'avons pas parallélisé d'impression des résultats. C'est pour deux raisons:
Tout comme dans le cas de detach_task() contre submit_task() , vous pouvez parfois paralléliser une boucle, mais vous n'en avez pas besoin pour retourner un BS::multi_future . Dans ce cas, vous pouvez enregistrer la surcharge de la génération des contrats à terme (qui peuvent être significatifs, en fonction du nombre de blocs) en utilisant detach_loop() au lieu de submit_loop() , avec les mêmes arguments.
Par exemple, nous pourrions détacher la boucle des carrés d'exemple ci-dessus comme suit:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iomanip > // std::setw
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool ( 10 );
constexpr unsigned int max = 100 ;
unsigned int squares[max];
pool. detach_loop < unsigned int >( 0 , max,
[&squares]( const unsigned int i)
{
squares[i] = i * i;
});
pool. wait ();
for ( unsigned int i = 0 ; i < max; ++i)
std::cout << std::setw ( 2 ) << i << " ^2 = " << std::setw ( 4 ) << squares[i] << ((i % 5 != 4 ) ? " | " : " n " );
} Warning: Since detach_loop() does not return a BS::multi_future , there is no built-in way for the user to know when the loop finishes executing. You must use either wait() as we did here, or some other method such as condition variables, to ensure that the loop finishes executing before trying to use anything that depends on its output. Otherwise, bad things will happen!
We have seen that detach_loop() and submit_loop() execute the function loop(i) for each index i in the loop. However, behind the scenes, the loop is split into blocks, and each block executes the loop() function multiple times. Each block has an internal loop of the form (where T is the type of the indices):
for (T i = start; i < end; ++i)
loop (i); The start and end indices of each block are determined automatically by the pool. For example, in the previous section, the loop from 0 to 100 was split into 10 blocks of 10 indices each: start = 0 to end = 10 , start = 10 to end = 20 , and so on; the blocks are not inclusive of the last index, since the for loop has the condition i < end and not i <= end .
However, this also means that the loop() function is executed multiple times per block. This generates additional overhead due to the multiple function calls. For short loops, this should not affect performance. However, for very long loops, with millions of indices, the performance cost may be significate.
For this reason, the thread pool library provides two additional member functions for parallelizing loops: detach_blocks() and submit_blocks() . While detach_loop() and submit_loop() execute a function loop(i) once per index but multiple times per block, detach_blocks() and submit_blocks() execute a function block(start, end) once per block.
The main advantage of this method is increased performance, but the main disadvantage is slightly more complicated code. In particular, the user must define the loop from start to end manually within each block. Here is the previous example using detach_blocks() :
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iomanip > // std::setw
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool ( 10 );
constexpr unsigned int max = 100 ;
unsigned int squares[max];
pool. detach_blocks < unsigned int >( 0 , max,
[&squares]( const unsigned int start, const unsigned int end)
{
for ( unsigned int i = start; i < end; ++i)
squares[i] = i * i;
});
pool. wait ();
for ( unsigned int i = 0 ; i < max; ++i)
std::cout << std::setw ( 2 ) << i << " ^2 = " << std::setw ( 4 ) << squares[i] << ((i % 5 != 4 ) ? " | " : " n " );
}Note how the block function takes two arguments, and includes the internal loop.
Generally, compiler optimizations should be able to make detach_loop() and submit_loop() perform roughly the same as detach_blocks() and submit_blocks() . However, you should perform your own benchmarks to see which option works best for your particular use case.
Unlike submit_task() , the member function submit_loop() only takes loop functions with no return values. The reason is that it wouldn't make sense to return a future for every single index of the loop. However, submit_blocks() does allow the block function to have a return value, as the number of blocks will generally not be too large, unlike the number of indices.
The block function will be executed once for each block, but the blocks are managed by the thread pool, with the user only able to select the number of blocks, but not the range of each block. Therefore, there is limited usability in returning one value per block. However, for cases where this is desired, such as for summation or some sorting algorithms, submit_blocks() does accept functions with return values, in which case it returns a BS::multi_future<T> object where T is the type of the return values.
Here's an example of a function template summing all elements of type T in a given range:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < cstdint > // std::uint64_t
# include < future > // std::future
# include < iostream > // std::cout
BS::thread_pool pool;
template < typename T>
T sum (T min, T max)
{
BS::multi_future<T> loop_future = pool. submit_blocks <T>(
min, max + 1 ,
[]( const T start, const T end)
{
T block_total = 0 ;
for (T i = start; i < end; ++i)
block_total += i;
return block_total;
},
100 );
T result = 0 ;
for (std::future<T>& future : loop_future)
result += future. get ();
return result;
}
int main ()
{
std::cout << sum<std:: uint64_t >( 1 , 1'000'000 );
} Here we used the fact that BS::multi_future<T> is a specialization of std::vector<std::future<T>> , so we can use a range-based for loop to iterate over the futures, and use the get() member function of each future to get its value. The values of the futures will be the partial sums from each block, so when we add them up, we will get the total sum. Note that we divided the loop into 100 blocks, so there will be 100 futures in total, each with the partial sum of 10,000 numbers.
The range-based for loop will likely start before the loop finished executing, and each time it calls a future, it will get the value of that future if it is ready, or it will wait until the future is ready and then get the value. This increases performance, since we can start summing the results without waiting for the entire loop to finish executing first - we only need to wait for individual blocks.
If we did want to wait until the entire loop finishes before summing the results, we could have used the get() member function of the BS::multi_future<T> object itself, which returns an std::vector<T> with the values obtained from each future. In that case, the sum could be obtained after calling submit_blocks() as follows:
std::vector<T> partial_sums = loop_future.get();
T result = std::reduce(partial_sums.begin(), partial_sums.end());
return result; The member functions detach_loop() , submit_loop() , detach_blocks() , and submit_blocks() parallelize a loop by splitting it into blocks, and submitting each block as an individual task to the queue, with each such task iterating over all the indices in the corresponding block's range, which can be numerous. However, sometimes we have loops with few indices, or more generally, a sequence of tasks enumerated by some index. In such cases, we can avoid the overhead of splitting into blocks and simply submit each individual index as its own independent task to the pool's queue.
This can be done with detach_sequence() and submit_sequence() . The syntax of these functions is similar to detach_loop() and submit_loop() , except that they don't have the num_blocks argument at the end. The sequence function must take only one argument, the index. As usual, detach_sequence() detaches the tasks and does not return a future, while submit_sequence() returns a BS::multi_future . If the tasks in the sequence return values, then the futures will contain those values, otherwise they will be void futures.
Here is a simple example:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < cstdint > // std::uint64_t
# include < iostream > // std::cout
# include < vector > // std::vector
using ui64 = std:: uint64_t ;
ui64 factorial ( const ui64 n)
{
ui64 result = 1 ;
for (ui64 i = 2 ; i <= n; ++i)
result *= i;
return result;
}
int main ()
{
BS::thread_pool pool;
constexpr ui64 max = 20 ;
BS::multi_future<ui64> sequence_future = pool. submit_sequence <ui64>( 0 , max + 1 , factorial);
std::vector<ui64> factorials = sequence_future. get ();
for (ui64 i = 0 ; i < max + 1 ; ++i)
std::cout << i << " ! = " << factorials[i] << ' n ' ;
}BS::multi_future<T> The helper class template BS::multi_future<T> , which we have been using throughout this section, provides a convenient way to collect and access groups of futures. This class is a specialization of std::vector<T> , so it should be used in a similar way:
[] operator to access the future at a specific index, or the push_back() member function to append a new future to the list.size() member function tells you how many futures are currently stored in the object. However, BS::multi_future<T> also has additional member functions that are aimed specifically at handling futures:
wait() to wait for all of them at once or get() to get an std::vector<T> with the results from all of them.ready_count() .valid() .wait_for() or wait until a specific time with wait_until() . These functions return true if all futures have been waited for before the duration expired or the time point was reached, and false otherwise. Aside from using BS::multi_future<T> to track the execution of parallelized loops, it can also be used, for example, whenever you have several different groups of tasks and you want to track the execution of each group individually.
The optional header file BS_thread_pool_utils.hpp contains several useful utility classes. These are not necessary for using the thread pool itself; BS_thread_pool.hpp is the only header file required. However, the utility classes can make writing multithreading code more convenient.
As with the main header file, the version of the utilities header file can be found by checking three macros:
BS_THREAD_POOL_UTILS_VERSION_MAJOR - indicates the major version.BS_THREAD_POOL_UTILS_VERSION_MINOR - indicates the minor version.BS_THREAD_POOL_UTILS_VERSION_PATCH - indicates the patch version.BS::synced_streamWhen printing to an output stream from multiple threads in parallel, the output may become garbled. For example, consider this code:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iostream > // std::cout
BS::thread_pool pool;
int main ()
{
pool. detach_sequence ( 0 , 5 ,
[]( int i)
{
std::cout << " Task no. " << i << " executing. n " ;
});
}The output will be a mess similar to this:
Task no. Task no. Task no. 3 executing.
0 executing.
Task no. 41 executing.
Task no. 2 executing.
executing.
The reason is that, although each individual insertion to std::cout is thread-safe, there is no mechanism in place to ensure subsequent insertions from the same thread are printed contiguously.
The utility class BS::synced_stream is designed to eliminate such synchronization issues. The constructor takes one optional argument, specifying the output stream to print to. If no argument is supplied, std::cout will be used:
// Construct a synced stream that will print to std::cout.
BS::synced_stream sync_out;
// Construct a synced stream that will print to the output stream my_stream.
BS::synced_stream sync_out (my_stream); The member function print() takes an arbitrary number of arguments, which are inserted into the stream one by one, in the order they were given. println() does the same, but also prints a newline character n at the end, for convenience. A mutex is used to synchronize this process, so that any other calls to print() or println() using the same BS::synced_stream object must wait until the previous call has finished.
As an example, this code:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
BS::synced_stream sync_out;
BS::thread_pool pool;
int main ()
{
pool. detach_sequence ( 0 , 5 ,
[]( int i)
{
sync_out. println ( " Task no. " , i, " executing. " );
});
}Will print out:
Task no. 0 executing.
Task no. 1 executing.
Task no. 2 executing.
Task no. 3 executing.
Task no. 4 executing.
Warning: Always create the BS::synced_stream object before the BS::thread_pool object, as we did in this example. When the BS::thread_pool object goes out of scope, it waits for the remaining tasks to be executed. If the BS::synced_stream object goes out of scope before the BS::thread_pool object, then any tasks using the BS::synced_stream will crash. Since objects are destructed in the opposite order of construction, creating the BS::synced_stream object before the BS::thread_pool object ensures that the BS::synced_stream is always available to the tasks, even while the pool is destructing.
Most stream manipulators defined in the headers <ios> and <iomanip> , such as std::setw (set the character width of the next output), std::setprecision (set the precision of floating point numbers), and std::fixed (display floating point numbers with a fixed number of digits), can be passed to print() and println() just as you would pass them to a stream.
The only exceptions are the flushing manipulators std::endl and std::flush , which will not work because the compiler will not be able to figure out which template specializations to use. Instead, use BS::synced_stream::endl and BS::synced_stream::flush . Here is an example:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < cmath > // std::sqrt
# include < iomanip > // std::setprecision, std::setw
# include < ios > // std::fixed
BS::synced_stream sync_out;
BS::thread_pool pool;
int main ()
{
sync_out. print ( std::setprecision ( 10 ), std::fixed);
pool. detach_sequence ( 0 , 16 ,
[]( int i)
{
sync_out. print ( " The square root of " , std::setw ( 2 ), i, " is " , std::sqrt (i), " . " , BS::synced_stream::endl);
});
} Note, however, that BS::synced_stream::endl should only be used if flushing is desired; otherwise, a newline character should be used instead.
BS::timerIf you are using a thread pool, then your code is most likely performance-critical. Achieving maximum performance requires performing a considerable amount of benchmarking to determine the optimal settings and algorithms. Therefore, it is important to be able to measure the execution time of various computations and operations under different conditions.
The utility class BS::timer provides a simple way to measure execution time. It is very straightforward to use:
BS::timer object.start() member function.stop() member function.ms() to obtain the elapsed time for the computation in milliseconds.current_ms() to obtain the elapsed time so far but keep the timer ticking.Par exemple:
BS::timer tmr;
tmr.start();
do_something ();
tmr.stop();
std::cout << " The elapsed time was " << tmr.ms() << " ms. n " ; A practical application of the BS::timer class can be found in the benchmark portion of the test program BS_thread_pool_test.cpp .
BS::signaller BS::signaller is a utility class which can be used to allow simple signalling between threads. To use it, construct an object and then pass it to the different threads. Multiple threads can call the wait() member function of the signaller. When another thread calls the ready() member function, the waiting threads will stop waiting.
That's really all there is to it; BS::signaller is really just a convenient wrapper around std::promise , which contains both the promise and its future. For usage examples, please see the test program BS_thread_pool_test.cpp .
Sometimes you may wish to monitor what is happening with the tasks you submitted to the pool. This may be done using three member functions:
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.get_tasks_total() == get_tasks_queued() + get_tasks_running() .These functions are demonstrated in the following program:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < chrono > // std::chrono
# include < thread > // std::this_thread
BS::synced_stream sync_out;
BS::thread_pool pool ( 4 );
void sleep_half_second ( const int i)
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 500 ));
sync_out. println ( " Task " , i, " done. " );
}
void monitor_tasks ()
{
sync_out. println (pool. get_tasks_total (), " tasks total, " , pool. get_tasks_running (), " tasks running, " , pool. get_tasks_queued (), " tasks queued. " );
}
int main ()
{
pool. wait ();
pool. detach_sequence ( 0 , 12 , sleep_half_second);
monitor_tasks ();
std::this_thread::sleep_for ( std::chrono::milliseconds ( 750 ));
monitor_tasks ();
std::this_thread::sleep_for ( std::chrono::milliseconds ( 500 ));
monitor_tasks ();
std::this_thread::sleep_for ( std::chrono::milliseconds ( 500 ));
monitor_tasks ();
}Assuming you have at least 4 hardware threads (so that 4 tasks can run concurrently), the output should be similar to:
12 tasks total, 0 tasks running, 12 tasks queued.
Task 0 done.
Task 1 done.
Task 2 done.
Task 3 done.
8 tasks total, 4 tasks running, 4 tasks queued.
Task 4 done.
Task 5 done.
Task 6 done.
Task 7 done.
4 tasks total, 4 tasks running, 0 tasks queued.
Task 8 done.
Task 9 done.
Task 10 done.
Task 11 done.
0 tasks total, 0 tasks running, 0 tasks queued.
The reason we called pool.wait() in the beginning is that when the thread pool is created, an initialization task runs in each thread, so if we don't wait, the first line will say there are 16 tasks in total, including the 4 initialization tasks. Voir ci-dessous pour plus de détails.
Consider a situation where the user cancels a multithreaded operation while it is still ongoing. Perhaps the operation was split into multiple tasks, and half of the tasks are currently being executed by the pool's threads, but the other half are still waiting in the queue.
The thread pool cannot terminate the tasks that are already running, as the C++17 standard does not provide that functionality (and in any case, abruptly terminating a task while it's running could have extremely bad consequences, such as memory leaks and data corruption). However, the tasks that are still waiting in the queue can be purged using the purge() member function.
Once purge() is called, any tasks still waiting in the queue will be discarded, and will never be executed by the threads. Please note that there is no way to restore the purged tasks; they are gone forever.
Consider for example the following program:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < chrono > // std::chrono
# include < thread > // std::this_thread
BS::synced_stream sync_out;
BS::thread_pool pool ( 4 );
int main ()
{
for ( size_t i = 0 ; i < 8 ; ++i)
{
pool. detach_task (
[i]
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 100 ));
sync_out. println ( " Task " , i, " done. " );
});
}
std::this_thread::sleep_for ( std::chrono::milliseconds ( 50 ));
pool. purge ();
pool. wait ();
} The program submit 8 tasks to the queue. Each task waits 100 milliseconds and then prints a message. The thread pool has 4 threads, so it will execute the first 4 tasks in parallel, and then the remaining 4. We wait 50 milliseconds, to ensure that the first 4 tasks have all started running. Then we call purge() to purge the remaining 4 tasks. As a result, these tasks never get executed. However, since the first 4 tasks are still running when purge() is called, they will finish uninterrupted; purge() only discards tasks that have not yet started running. The output of the program therefore only contains the messages from the first 4 tasks:
Task 0 done.
Task 1 done.
Task 2 done.
Task 3 done.
submit_task() catches any exceptions thrown by the submitted task and forwards them to the corresponding future. They can then be caught when invoking the get() member function of the future. Par exemple:
# include " BS_thread_pool.hpp "
BS::synced_stream sync_out;
BS::thread_pool pool;
double inverse ( const double x)
{
if (x == 0 )
throw std::runtime_error ( " Division by zero! " );
else
return 1 / x;
}
int main ()
{
constexpr double num = 0 ;
std::future< double > my_future = pool. submit_task (inverse, num);
try
{
const double result = my_future. get ();
sync_out. println ( " The inverse of " , num, " is " , result, " . " );
}
catch ( const std:: exception & e)
{
sync_out. println ( " Caught exception: " , e. what ());
}
}The output will be:
Caught exception: Division by zero!
However, if you change num to any non-zero number, no exceptions will be thrown and the inverse will be printed.
It is important to note that wait() does not throw any exceptions; only get() does. Therefore, even if your task does not return anything, ie your future is an std::future<void> , you must still use get() on the future obtained from it if you want to catch exceptions thrown by it. Here is an example:
# include " BS_thread_pool.hpp "
BS::synced_stream sync_out;
BS::thread_pool pool;
void print_inverse ( const double x)
{
if (x == 0 )
throw std::runtime_error ( " Division by zero! " );
else
sync_out. println ( " The inverse of " , x, " is " , 1 / x, " . " );
}
int main ()
{
constexpr double num = 0 ;
std::future< void > my_future = pool. submit_task (print_inverse, num);
try
{
my_future. get ();
}
catch ( const std:: exception & e)
{
sync_out. println ( " Caught exception: " , e. what ());
}
} When using BS::multi_future to handle multiple futures at once, exception handling works the same way: if any of the futures may throw exceptions, you may catch these exceptions when calling get() , even in the case of BS::multi_future<void> .
If you do not require exception handling, or if exceptions are explicitly disabled in your codebase, you can define the macro BS_THREAD_POOL_DISABLE_EXCEPTION_HANDLING before including BS_thread_pool.hpp , which will disable exception handling in submit_task() . Note that if the feature-test macro __cpp_exceptions is undefined, BS_THREAD_POOL_DISABLE_EXCEPTION_HANDLING will be automatically defined.
BS::thread_pool comes with a variety of methods to obtain information about the threads in the pool:
BS::this_thread provides functionality similar to std::this_thread . If the current thread belongs to a BS::thread_pool object, then BS::this_thread::get_index() can be used to get the index of the current thread, and BS::this_thread::get_pool() can be used to get the pointer to the thread pool that owns the current thread. Please see the reference below for more details.get_thread_ids() returns a vector containing the unique identifiers for each of the pool's threads, as obtained by std::thread::get_id() . These values are not so useful on their own, but can be used for whatever the user wants to use them for.get_native_handles() , if enabled, returns a vector containing the underlying implementation-defined thread handles for each of the pool's threads, as obtained by std::thread::native_handle() . For more information, see the relevant section below. Sometimes, it is necessary to initialize the threads before they run any tasks. This can be done by submitting a proper initialization function to the constructor or to reset() , either as the only argument or as the second argument after the desired number of threads. The thread initialization must take no arguments and have no return value. However, if needed, the function can use BS::this_thread::get_index() and BS::this_thread::get_pool() to figure out which thread and pool it belongs to.
The thread initialization function is submitted as a set of special tasks, one per thread, which bypass the queue, but still count towards the number of running tasks, which means get_tasks_total() and get_tasks_running() will report that these tasks are running if they are checked immediately after the pool is initialized.
This is done so that the user has the option to either wait for the initialization tasks to finish, by calling wait() on the pool, or just keep going. In either case, the initialization tasks will always finish executing before any tasks are picked out of the queue, so there is no reason to wait for them to finish unless they have some side-effects that affect the main thread.
Here is a simple example:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < random > // std::mt19937_64, std::random_device
BS::synced_stream sync_out;
thread_local std::mt19937_64 twister;
int main ()
{
BS::thread_pool pool (
[]
{
twister. seed ( std::random_device ()());
});
pool. submit_sequence ( 0 , 4 ,
[]( int )
{
sync_out. println ( " I generated a random number: " , twister ());
})
. wait ();
} In this example, we create a thread_local Mersenne twister engine, meaning that each thread has its own independent engine. However, we did not seed the engine, so each thread will generate the exact same sequence of pseudo-random numbers. To remedy this, we pass an initialization function to the BS::thread_pool constructor which seeds the twister in each thread with the (hopefully) non-deterministic random number generator std::random_device .
In C++, it is often crucial to pass function arguments by reference or constant reference, instead of by value. This allows the function to access the object being passed directly, rather than creating a new copy of the object. We have already seen that submitting an argument by reference is a simple matter of capturing it with a & in the lambda capture list. To submit as constant reference, we can use std::as_const as in the following example:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < utility > // std::as_const
BS::synced_stream sync_out;
void increment ( int & x)
{
++x;
}
void print ( const int & x)
{
sync_out. println (x);
}
int main ()
{
BS::thread_pool pool;
int n = 0 ;
pool. submit_task (
[&n]
{
increment (n);
})
. wait ();
pool. submit_task (
[&n = std::as_const (n)]
{
print (n);
})
. wait ();
} The increment() function takes a reference to an integer, and increments that integer. Passing the argument by reference guarantees that n itself, in the scope of main() , will be incremented - rather than a copy of it in the scope of increment() .
Similarly, the print() function takes a constant reference to an integer, and prints that integer. Passing the argument by constant reference guarantees that the variable will not be accidentally modified by the function, even though we are accessing n itself, rather than a copy. If we replace print with increment , the program won't compile, as increment cannot take constant references.
Generally, it is not really necessary to pass arguments by constant reference, but it is more "correct" to do so, if we would like to guarantee that the variable being referenced is indeed never modified. This section is therefore included here for completeness.
Sometimes you may wish to temporarily pause the execution of tasks, or perhaps you want to submit tasks to the queue in advance and only start executing them at a later time. You can do this using the member functions pause() , unpause() , and is_paused() .
However, these functions are disabled by default, and must be explicitly enabled by defining the macro BS_THREAD_POOL_ENABLE_PAUSE before including BS_thread_pool.hpp . The reason is that pausing the pool adds additional checks to the waiting and worker functions, which have a very small but non-zero overhead.
When you call pause() , the workers will temporarily stop retrieving new tasks out of the queue. However, any tasks already executed will keep running until they are done, since the thread pool has no control over the internal code of your tasks. If you need to pause a task in the middle of its execution, you must do that manually by programming your own pause mechanism into the task itself. To resume retrieving tasks, call unpause() . To check whether the pool is currently paused, call is_paused() .
Here is an example:
# define BS_THREAD_POOL_ENABLE_PAUSE
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < chrono > // std::chrono
# include < thread > // std::this_thread
BS::synced_stream sync_out;
BS::thread_pool pool ( 4 );
void sleep_half_second ( const int i)
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 500 ));
sync_out. println ( " Task " , i, " done. " );
}
void check_if_paused ()
{
if (pool. is_paused ())
sync_out. println ( " Pool paused. " );
else
sync_out. println ( " Pool unpaused. " );
}
int main ()
{
pool. detach_sequence ( 0 , 8 , sleep_half_second);
sync_out. println ( " Submitted 8 tasks. " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 250 ));
pool. pause ();
check_if_paused ();
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
sync_out. println ( " Still paused... " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
pool. detach_sequence ( 8 , 12 , sleep_half_second);
sync_out. println ( " Submitted 4 more tasks. " );
sync_out. println ( " Still paused... " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
pool. unpause ();
check_if_paused ();
}Assuming you have at least 4 hardware threads, the output should be similar to:
Submitted 8 tasks.
Pool paused.
Task 0 done.
Task 1 done.
Task 2 done.
Task 3 done.
Still paused...
Submitted 4 more tasks.
Still paused...
Pool unpaused.
Task 4 done.
Task 5 done.
Task 6 done.
Task 7 done.
Task 8 done.
Task 9 done.
Task 10 done.
Task 11 done.
Here is what happened. We initially submitted a total of 8 tasks to the queue. Since we waited for 250ms before pausing, the first 4 tasks have already started running, so they kept running until they finished. While the pool was paused, we submitted 4 more tasks to the queue, but they just waited at the end of the queue. When we unpaused, the remaining 4 initial tasks were executed, followed by the 4 new tasks.
While the workers are paused, wait() will wait for the running tasks instead of all tasks (otherwise it would wait forever). This is demonstrated by the following program:
# define BS_THREAD_POOL_ENABLE_PAUSE
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < chrono > // std::chrono
# include < thread > // std::this_thread
BS::synced_stream sync_out;
BS::thread_pool pool ( 4 );
void sleep_half_second ( const int i)
{
std::this_thread::sleep_for ( std::chrono::milliseconds ( 500 ));
sync_out. println ( " Task " , i, " done. " );
}
void check_if_paused ()
{
if (pool. is_paused ())
sync_out. println ( " Pool paused. " );
else
sync_out. println ( " Pool unpaused. " );
}
int main ()
{
pool. detach_sequence ( 0 , 8 , sleep_half_second);
sync_out. println ( " Submitted 8 tasks. Waiting for them to complete. " );
pool. wait ();
pool. detach_sequence ( 8 , 20 , sleep_half_second);
sync_out. println ( " Submitted 12 more tasks. " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 250 ));
pool. pause ();
check_if_paused ();
sync_out. println ( " Waiting for the " , pool. get_tasks_running (), " running tasks to complete. " );
pool. wait ();
sync_out. println ( " All running tasks completed. " , pool. get_tasks_queued (), " tasks still queued. " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
sync_out. println ( " Still paused... " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
sync_out. println ( " Still paused... " );
std::this_thread::sleep_for ( std::chrono::milliseconds ( 1000 ));
pool. unpause ();
check_if_paused ();
std::this_thread::sleep_for ( std::chrono::milliseconds ( 250 ));
sync_out. println ( " Waiting for the remaining " , pool. get_tasks_total (), " tasks ( " , pool. get_tasks_running (), " running and " , pool. get_tasks_queued (), " queued) to complete. " );
pool. wait ();
sync_out. println ( " All tasks completed. " );
}The output should be similar to:
Submitted 8 tasks. Waiting for them to complete.
Task 0 done.
Task 1 done.
Task 2 done.
Task 3 done.
Task 4 done.
Task 5 done.
Task 6 done.
Task 7 done.
Submitted 12 more tasks.
Pool paused.
Waiting for the 4 running tasks to complete.
Task 8 done.
Task 9 done.
Task 10 done.
Task 11 done.
All running tasks completed. 8 tasks still queued.
Still paused...
Still paused...
Pool unpaused.
Waiting for the remaining 8 tasks (4 running and 4 queued) to complete.
Task 12 done.
Task 13 done.
Task 14 done.
Task 15 done.
Task 16 done.
Task 17 done.
Task 18 done.
Task 19 done.
All tasks completed.
The first wait() , which was called while the pool was not paused, waited for all 8 tasks, both running and queued. The second wait() , which was called after pausing the pool, only waited for the 4 running tasks, while the other 8 tasks remained queued, and were not executed since the pool was paused. Finally, the third wait() , which was called after unpausing the pool, waited for the remaining 8 tasks, both running and queued.
Warning: If the thread pool is destroyed while paused, any tasks still in the queue will never be executed!
Consider the following program:
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool;
pool. detach_task (
[&pool]
{
pool. wait ();
std::cout << " Done waiting. n " ;
});
}This program creates a thread pool, and then detaches a task that waits for tasks in the same thread pool to complete. If you run this program, it will never print the message "Done waiting", because the task will wait for itself to complete. This causes a deadlock , and the program will wait forever.
Usually, in simple programs, this will never happen. However, in more complicated programs, perhaps ones running multiple thread pools in parallel, wait deadlocks could potentially occur. In such cases, the macro BS_THREAD_POOL_ENABLE_WAIT_DEADLOCK_CHECK can be defined before including BS_thread_pool.hpp . wait() will then check whether the user tried to call it from within a thread of the same pool, and if so, it will throw the exception BS::thread_pool::wait_deadlock instead of waiting. This check is disabled by default because wait deadlocks are not something that happens often, and the check adds a small but non-zero overhead every time wait() is called.
Here is an example:
# define BS_THREAD_POOL_ENABLE_WAIT_DEADLOCK_CHECK
# include " BS_thread_pool.hpp " // BS::thread_pool
# include < iostream > // std::cout
int main ()
{
BS::thread_pool pool;
pool. detach_task (
[&pool]
{
try
{
pool. wait ();
std::cout << " Done waiting. n " ;
}
catch ( const BS::thread_pool::wait_deadlock&)
{
std::cout << " Error: Deadlock! n " ;
}
});
} This time, wait() will detect the deadlock, and will throw an exception, causing the output to be "Error: Deadlock!" .
Note that if the feature-test macro __cpp_exceptions is undefined, BS_THREAD_POOL_ENABLE_WAIT_DEADLOCK_CHECK will be automatically undefined.
The BS::thread_pool member function get_native_handles() returns a vector containing the underlying implementation-defined thread handles for each of the pool's threads. These can then be used in an implementation-specific way to manage the threads at the OS level
However, note that this will generally not be portable code. Furthermore, this feature uses std::thread::native_handle(), which is in the C++ standard library, but is not guaranteed to be present on all systems. Therefore, this feature is turned off by default, and must be turned on by defining the macro BS_THREAD_POOL_ENABLE_NATIVE_HANDLES before including BS_thread_pool.hpp .
Here is an example:
# define BS_THREAD_POOL_ENABLE_NATIVE_HANDLES
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
# include < thread > // std::thread
# include < vector > // std::vector
BS::synced_stream sync_out;
BS::thread_pool pool ( 4 );
int main ()
{
std::vector<std::thread::native_handle_type> handles = pool. get_native_handles ();
for (BS:: concurrency_t i = 0 ; i < handles. size (); ++i)
sync_out. println ( " Thread " , i, " native handle: " , handles[i]);
}The output will depend on your compiler and operating system. Here is an example:
Thread 0 native handle: 000000F4
Thread 1 native handle: 000000F8
Thread 2 native handle: 000000EC
Thread 3 native handle: 000000FC
Defining the macro BS_THREAD_POOL_ENABLE_PRIORITY before including BS_thread_pool.hpp enables task priority. The priority of a task or group of tasks may then be specified as an additional argument (at the end of the argument list) to detach_task() , submit_task() , detach_blocks() , submit_blocks() , detach_loop() , submit_loop() , detach_sequence() , and submit_sequence() . If the priority is not specified, the default value will be 0.
The priority is a number of type BS::priority_t , which is a signed 16-bit integer, so it can have any value between -32,768 and 32,767. The tasks will be executed in priority order from highest to lowest. If priority is assigned to the block/loop/sequence parallelization functions, which submit multiple tasks, then all of these tasks will have the same priority.
The namespace BS::pr contains some pre-defined priorities for users who wish to avoid magic numbers and enjoy better future-proofing. In order of decreasing priority, the pre-defined priorities are: BS::pr::highest , BS::pr::high , BS::pr::normal , BS::pr::low , and BS::pr::lowest .
Here is a simple example:
# define BS_THREAD_POOL_ENABLE_PRIORITY
# include " BS_thread_pool.hpp " // BS::thread_pool
# include " BS_thread_pool_utils.hpp " // BS::synced_stream
BS::synced_stream sync_out;
BS::thread_pool pool ( 1 );
int main ()
{
pool. detach_task ([] { sync_out. println ( " This task will execute third. " ); }, BS::pr:: normal );
pool. detach_task ([] { sync_out. println ( " This task will execute fifth. " ); }, BS::pr::lowest);
pool. detach_task ([] { sync_out. println ( " This task will execute second. " ); }, BS::pr::high);
pool. detach_task ([] { sync_out. println ( " This task will execute first. " ); }, BS::pr::highest);
pool. detach_task ([] { sync_out. println ( " This task will execute fourth. " ); }, BS::pr::low);
}This program will print out the tasks in the correct priority order. Note that for simplicity, we used a pool with just one thread, so the tasks will run one at a time. In a pool with 5 or more threads, all 5 tasks will actually run more or less at the same time, because, for example, the task with the second-highest priority will be picked up by another thread while the task with the highest priority is still running.
Of course, this is just a pedagogical example. In a realistic use case we may want, for example, to submit tasks that must be completed immediately with high priority so they skip over other tasks already in the queue, or background non-urgent tasks with low priority so they evaluate only after higher-priority tasks are done.
Here are some subtleties to note when using task priority:
std::priority_queue , which has O(log n) complexity for storing new tasks, but only O(1) complexity for retrieving the next (ie highest-priority) task. This is in contrast with std::queue , used if priority is disabled, which both stores and retrieves with O(1) complexity.std::priority_queue as a binary heap, which means tasks are stored as a binary tree instead of sequentially. To execute tasks in submission order, give them monotonically decreasing priorities.BS::priority_t is defined to be ( std::int_least16_t ), since this type is guaranteed to be present on all systems, rather than std::int16_t , which is optional in the C++ standard. This means that on some exotic systems BS::priority_t may actually have more than 16 bits. However, the pre-defined priorities are 100% portable, and will always have the same values (eg: BS::pr::highest = 32767 ) regardless of the actual bit width. The file BS_thread_pool_test.cpp in the tests folder of the GitHub repository will perform automated tests of all aspects of the library. The output will be printed both to std::cout and to a file with the same name as the executable and the suffix -yyyy-mm-dd_hh.mm.ss.log based on the current date and time. In addition, the code is meant to serve as an extensive example of how to properly use the library.
Please make sure to:
BS_thread_pool_test.cpp with optimization flags enabled (eg -O3 on GCC / Clang or /O2 on MSVC).The test program also takes command line arguments for automation purposes:
help : Show a help message and exit. Any other arguments will be ignored.log : Create a log file.tests : Perform standard tests.deadlock Perform long deadlock tests.benchmarks : Perform benchmarks. If no options are entered, the default is: log tests benchmarks .
By default, the test program enables all the optional features by defining the suitable macros, so it can test them. However, if the macro BS_THREAD_POOL_LIGHT_TEST is defined during compilation, the optional features will not be tested.
A PowerShell script, BS_thread_pool_test.ps1 , is provided for your convenience in the tests folder to make running the test on multiple compilers and operating systems easier. Since it is written in PowerShell, it is fully portable and works on Windows, Linux, and macOS. The script will automatically detect if Clang, GCC, and/or MSVC are available, and compile the test program using each available compiler twice - with and without all the optional features. It will then run each compiled test program and report on any errors.
If any of the tests fail, please submit a bug report including the exact specifications of your system (OS, CPU, compiler, etc.) and the generated log file.
If all checks passed, BS_thread_pool_test.cpp performs simple benchmarks by filling a very large vector with values using detach_blocks() . The program decides what the size of the vector should be by testing how many elements are needed to reach a certain target duration when parallelizing using a number of blocks equal to the number of threads. This ensures that the test takes approximately the same amount of time on all systems, and is thus more consistent and portable.
Once the appropriate size of the vector has been determined, the program allocates the vector and fills it with values, calculated according to a fixed prescription. This operation is performed both single-threaded and multithreaded, with the multithreaded computation spread across multiple tasks submitted to the pool.
Several different multithreaded tests are performed, with the number of tasks either equal to, smaller than, or larger than the pool's thread count. Each test is repeated multiple times, with the run times averaged over all runs of the same test. The program keeps increasing the number of blocks by a factor of 2 until diminishing returns are encountered. The run times of the tests are compared, and the maximum speedup obtained is calculated.
As an example, here are the results of the benchmarks from a Digital Research Alliance of Canada node equipped with two 20-core / 40-thread Intel Xeon Gold 6148 CPUs (for a total of 40 cores and 80 threads), running CentOS Linux 7.9.2009. The tests were compiled using GCC v13.2.0 with the -O3 and -march=native flags. The output was as follows:
======================
Performing benchmarks:
======================
Using 80 threads.
Determining the number of elements to generate in order to achieve an approximate mean execution time of 50 ms with 80 tasks...
Each test will be repeated up to 30 times to collect reliable statistics.
Generating 27962000 elements:
[......]
Single-threaded, mean execution time was 2815.2 ms with standard deviation 3.5 ms.
[......]
With 2 tasks, mean execution time was 1431.3 ms with standard deviation 10.1 ms.
[.......]
With 4 tasks, mean execution time was 722.1 ms with standard deviation 11.4 ms.
[..............]
With 8 tasks, mean execution time was 364.9 ms with standard deviation 10.9 ms.
[............................]
With 16 tasks, mean execution time was 181.9 ms with standard deviation 8.0 ms.
[..............................]
With 32 tasks, mean execution time was 110.6 ms with standard deviation 1.8 ms.
[..............................]
With 64 tasks, mean execution time was 64.0 ms with standard deviation 6.3 ms.
[..............................]
With 128 tasks, mean execution time was 59.8 ms with standard deviation 0.8 ms.
[..............................]
With 256 tasks, mean execution time was 59.0 ms with standard deviation 0.0 ms.
[..............................]
With 512 tasks, mean execution time was 52.8 ms with standard deviation 0.4 ms.
[..............................]
With 1024 tasks, mean execution time was 50.7 ms with standard deviation 0.9 ms.
[..............................]
With 2048 tasks, mean execution time was 50.0 ms with standard deviation 0.5 ms.
[..............................]
With 4096 tasks, mean execution time was 49.4 ms with standard deviation 0.5 ms.
[..............................]
With 8192 tasks, mean execution time was 50.2 ms with standard deviation 0.4 ms.
Maximum speedup obtained by multithreading vs. single-threading: 56.9x, using 4096 tasks.
+++++++++++++++++++++++++++++++++++++++
Thread pool performance test completed!
+++++++++++++++++++++++++++++++++++++++
These two CPUs have 40 physical cores in total, with each core providing two separate logical cores via hyperthreading, for a total of 80 threads. Without hyperthreading, we would expect a maximum theoretical speedup of 40x. With hyperthreading, one might naively expect to achieve up to an 80x speedup, but this is in fact impossible, as each pair of hyperthreaded logical cores share the same physical core's resources. However, generally we would expect at most an estimated 30% additional speedup from hyperthreading, which amounts to around 52x in this case. The speedup of 56.9x in our performance test exceeds this estimate.
If you are using the vcpkg C/C++ package manager, you can easily install BS::thread_pool with the following commands:
On Linux/macOS:
./vcpkg install bshoshany-thread-pool
On Windows:
.vcpkg install bshoshany-thread-pool:x86-windows bshoshany-thread-pool:x64-windows
To update the package to the latest version, run:
vcpkg upgrade
If you are using the Conan C/C++ package manager, you can easily integrate BS::thread_pool into your project by adding the following lines to your conanfile.txt :
[requires]
bshoshany-thread-pool/4.1.0To update the package to the latest version, simply change the version number. Please refer to this package's page on ConanCenter for more information.
If you are using the Meson build system, you can install BS::thread_pool from WrapDB. To do so, create a subprojects folder in your project (if it does not already exist) and run the following command:
meson wrap install bshoshany-thread-pool
Then, use dependency('bshoshany-thread-pool') in your meson.build file to include the package. To update the package to the latest version, run:
meson wrap update bshoshany-thread-pool
If you are using CMake, you can install BS::thread_pool with CPM. If CPM is already installed, simply add the following to your project's CMakeLists.txt :
CPMAddPackage(
NAME BS_thread_pool
GITHUB_REPOSITORY bshoshany/thread-pool
VERSION 4.1.0)
add_library (BS_thread_pool INTERFACE )
target_include_directories (BS_thread_pool INTERFACE ${BS_thread_pool_SOURCE_DIR} / include )This will automatically download the indicated version of the package from the GitHub repository and include it in your project.
It is also possible to use CPM without installing it first, by adding the following lines to CMakeLists.txt before CPMAddPackage :
set (CPM_DOWNLOAD_LOCATION " ${CMAKE_BINARY_DIR} /cmake/CPM.cmake" )
if ( NOT ( EXISTS ${CPM_DOWNLOAD_LOCATION} ))
message ( STATUS "Downloading CPM.cmake" )
file (DOWNLOAD https://github.com/cpm-cmake/CPM.cmake/releases/latest/download/CPM.cmake ${CPM_DOWNLOAD_LOCATION} )
endif ()
include ( ${CPM_DOWNLOAD_LOCATION} ) Here is an example of a complete CMakeLists.txt for a project named my_project consisting of a single source file main.cpp which uses BS_thread_pool.hpp :
cmake_minimum_required ( VERSION 3.19)
project (my_project LANGUAGES CXX)
set (CMAKE_CXX_STANDARD 17)
set (CMAKE_CXX_STANDARD_REQUIRED ON )
set (CMAKE_CXX_EXTENSIONS OFF )
set (CPM_DOWNLOAD_LOCATION " ${CMAKE_BINARY_DIR} /cmake/CPM.cmake" )
if ( NOT ( EXISTS ${CPM_DOWNLOAD_LOCATION} ))
message ( STATUS "Downloading CPM.cmake" )
file (DOWNLOAD https://github.com/cpm-cmake/CPM.cmake/releases/latest/download/CPM.cmake ${CPM_DOWNLOAD_LOCATION} )
endif ()
include ( ${CPM_DOWNLOAD_LOCATION} )
CPMAddPackage(
NAME BS_thread_pool
GITHUB_REPOSITORY bshoshany/thread-pool
VERSION 4.1.0)
add_library (BS_thread_pool INTERFACE )
target_include_directories (BS_thread_pool INTERFACE ${BS_thread_pool_SOURCE_DIR} / include )
add_executable (my_project main.cpp)
target_link_libraries (my_project BS_thread_pool) With both CMakeLists.txt and main.cpp in the same folder, type the following commands to build the project:
cmake -S . -B build
cmake --build build
This section provides a complete reference to classes, member functions, objects, and macros available in this library, along with other important information. Member functions are given here with simplified prototypes (eg removing const ) for ease of reading.
More information can be found in the provided Doxygen comments. Any modern IDE, such as Visual Studio Code, can use the Doxygen comments to provide automatic documentation for any class and member function in this library when hovering over code with the mouse or using auto-complete.
BS_thread_pool.hpp ) BS::thread_pool class The class BS::thread_pool is the main thread pool class. It can be used to create a pool of threads and submit tasks to a queue. When a thread becomes available, it takes a task from the queue and executes it. The member functions that are available by default, when no macros are defined, are:
thread_pool() : Construct a new thread pool. The number of threads will be 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.thread_pool(BS::concurrency_t num_threads) : Construct a new thread pool with the specified number of threads.thread_pool(std::function<void()>& init_task) : Construct a new thread pool with the specified initialization function.thread_pool(BS::concurrency_t num_threads, std::function<void()>& init_task) : Construct a new thread pool with the specified number of threads and initialization function.void reset() : Reset the pool with the total number of hardware threads available, as reported by the implementation. 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.void reset(BS::concurrency_t num_threads) : Reset the pool with a new number of threads.void reset(std::function<void()>& init_task) Reset the pool with the total number of hardware threads available, as reported by the implementation, and a new initialization function.void reset(BS::concurrency_t num_threads, std::function<void()>& init_task) : Reset the pool with a new number of threads and a new initialization function.size_t get_tasks_queued() : Get the number of tasks currently waiting in the queue to be executed by the threads.size_t get_tasks_running() : Get the number of tasks currently being executed by the threads.size_t get_tasks_total() : Get the total number of unfinished tasks: either still waiting in the queue, or running in a thread. Note that get_tasks_total() == get_tasks_queued() + get_tasks_running() .BS::concurrency_t get_thread_count() : Get the number of threads in the pool.std::vector<std::thread::id> get_thread_ids() : Get a vector containing the unique identifiers for each of the pool's threads, as obtained by std::thread::get_id() .T and F are template parameters):void detach_task(F&& task) : Submit a function with no arguments and no return value into the task queue. To push a function with arguments, enclose it in a lambda expression. Does not return a future, so the user must use wait() or some other method to ensure that the task finishes executing, otherwise bad things will happen.void detach_blocks(T first_index, T index_after_last, F&& block, size_t num_blocks = 0) : Parallelize a loop by automatically splitting it into blocks and submitting each block separately to the queue. The block function takes two arguments, the start and end of the block, so that it is only called only once per block, but it is up to the user make sure the block function correctly deals with all the indices in each block. Does not return a BS::multi_future , so the user must use wait() or some other method to ensure that the loop finishes executing, otherwise bad things will happen.void detach_loop(T first_index, T index_after_last, F&& loop, size_t num_blocks = 0) : Parallelize a loop by automatically splitting it into blocks and submitting each block separately to the queue. The loop function takes one argument, the loop index, so that it is called many times per block. Does not return a BS::multi_future , so the user must use wait() or some other method to ensure that the loop finishes executing, otherwise bad things will happen.void detach_sequence(T first_index, T index_after_last, F&& sequence) : Submit a sequence of tasks enumerated by indices to the queue. Does not return a BS::multi_future , so the user must use wait() or some other method to ensure that the sequence finishes executing, otherwise bad things will happen.T , F , and R are template parameters):std::future<R> submit_task(F&& task) : Submit a function with no arguments into the task queue. To submit a function with arguments, enclose it in a lambda expression. 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.BS::multi_future<R> submit_blocks(T first_index, T index_after_last, F&& block, size_t num_blocks = 0) : Parallelize a loop by automatically splitting it into blocks and submitting each block separately to the queue. The block function takes two arguments, the start and end of the block, so that it is only called only once per block, but it is up to the user make sure the block function correctly deals with all the indices in each block. Returns a BS::multi_future that contains the futures for all of the blocks.BS::multi_future<void> submit_loop(T first_index, T index_after_last, F&& loop, size_t num_blocks = 0) : Parallelize a loop by automatically splitting it into blocks and submitting each block separately to the queue. The loop function takes one argument, the loop index, so that it is called many times per block. It must have no return value. Returns a BS::multi_future that contains the futures for all of the blocks.BS::multi_future<R> submit_sequence(T first_index, T index_after_last, F&& sequence) : Submit a sequence of tasks enumerated by indices to the queue. Returns a BS::multi_future that contains the futures for all of the tasks.void purge() : Purge all the tasks waiting in the queue. Tasks that are currently running will not be affected, but any tasks still waiting in the queue will be discarded, and will never be executed by the threads. Please note that there is no way to restore the purged tasks.R and P , C , and D are template parameters):void wait() : 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_task() instead, and call the wait() member function of the generated future.bool wait_for(std::chrono::duration<R, P>& duration) : Wait for tasks to be completed, but stop waiting after the specified duration has passed. Returns true if all tasks finished running, false if the duration expired but some tasks are still running.bool wait_until(std::chrono::time_point<C, D>& timeout_time) : Wait for tasks to be completed, but stop waiting after the specified time point has been reached. Returns true if all tasks finished running, false if the time point was reached but some tasks are still running. When a BS::thread_pool object goes out of scope, the destructor first waits for all tasks to complete, then destroys all threads. Note that if the pool is paused, then any tasks still in the queue will never be executed.
BS::thread_pool classThe thread pool has several optional features that must be explicitly enabled using macros.
BS_THREAD_POOL_ENABLE_PRIORITY .detach_task() , submit_task() , detach_blocks() , submit_blocks() , detach_loop() , submit_loop() , detach_sequence() , and submit_sequence() . If the priority is not specified, the default value will be 0.BS::priority_t , which is a signed 16-bit integer, so it can have any value between -32,768 and 32,767. The tasks will be executed in priority order from highest to lowest.BS::pr contains some pre-defined priorities: BS::pr::highest , BS::pr::high , BS::pr::normal , BS::pr::low , and BS::pr::lowest .BS_THREAD_POOL_ENABLE_PAUSE . Adds the following member functions:void pause() : Pause the pool. The workers will temporarily stop retrieving new tasks out of the queue, although any tasks already executed will keep running until they are finished.void unpause() : Unpause the pool. The workers will resume retrieving new tasks out of the queue.bool is_paused() : Check whether the pool is currently paused.BS_THREAD_POOL_ENABLE_NATIVE_HANDLES . Adds the following member function:std::vector<std::thread::native_handle_type> get_native_handles() : Get a vector containing the underlying implementation-defined thread handles for each of the pool's threads.BS_THREAD_POOL_ENABLE_WAIT_DEADLOCK_CHECK .wait() , wait_for() , and wait_until() will check whether the user tried to call them from within a thread of the same pool, which would result in a deadlock. If so, they will throw the exception BS::thread_pool::wait_deadlock instead of waiting.BS_THREAD_POOL_DISABLE_EXCEPTION_HANDLING .submit_task() if it is not needed, or if exceptions are explicitly disabled in the codebase.BS_THREAD_POOL_ENABLE_WAIT_DEADLOCK_CHECK . Disabling exception handling removes the try - catch block from submit_task() , while enabling wait deadlock checks adds a throw expression to wait() , wait_for() , and wait_until() .__cpp_exceptions is undefined, BS_THREAD_POOL_DISABLE_EXCEPTION_HANDLING is automatically defined, and BS_THREAD_POOL_ENABLE_WAIT_DEADLOCK_CHECK is automatically undefined. BS::this_thread namespace The namespace BS::this_thread provides functionality similar to std::this_thread . It contains the following function objects:
BS::this_thread::get_index() can be used to get the index of the current thread. If this thread belongs to a BS::thread_pool object, it will have an index from 0 to BS::thread_pool::get_thread_count() - 1 . Otherwise, for example if this thread is the main thread or an independent std::thread , std::nullopt will be returned.BS::this_thread::get_pool() can be used to get the pointer to the thread pool that owns the current thread. If this thread belongs to a BS::thread_pool object, a pointer to that object will be returned. Otherwise, std::nullopt will be returned.std::optional object will be returned, of type BS::this_thread::optional_index or BS::this_thread::optional_pool respectively. Unless you are 100% sure this thread is in a pool, first use std::optional::has_value() to check if it contains a value, and if so, use std::optional::value() to obtain that value. BS::multi_future<T> class BS::multi_future<T> is a helper class used to facilitate waiting for and/or getting the results of multiple futures at once. It is defined as a specialization of std::vector<std::future<T>> . This means that all of the member functions that can be used on an std::vector can also be used on a BS::multi_future . For example, you may use a range-based for loop with a BS::multi_future , since it has iterators.
In addition to inherited member functions, BS::multi_future has the following specialized member functions ( R and P , C , and D are template parameters):
[void or std::vector<T>] get() : Get the results from all the futures stored in this BS::multi_future , rethrowing any stored exceptions. If the futures return void , this function returns void as well. If the futures return a type T , this function returns a vector containing the results.size_t ready_count() : Check how many of the futures stored in this BS::multi_future are ready.bool valid() : Check if all the futures stored in this BS::multi_future are valid.void wait() : Wait for all the futures stored in this BS::multi_future .bool wait_for(std::chrono::duration<R, P>& duration) : Wait for all the futures stored in this BS::multi_future , but stop waiting after the specified duration has passed. Returns true if all futures have been waited for before the duration expired, false otherwise.bool wait_until(std::chrono::time_point<C, D>& timeout_time) : Wait for all the futures stored in this multi_future object, but stop waiting after the specified time point has been reached. Returns true if all futures have been waited for before the time point was reached, false otherwise.BS_thread_pool_utils.hpp ) BS::signaller class BS::signaller is a utility class which can be used to allow simple signalling between threads. This class is really just a convenient wrapper around std::promise , which contains both the promise and its future. It has the following member functions:
signaller() : Construct a new signaller.void wait() : Wait until the signaller is ready.void ready() : Inform any waiting threads that the signaller is ready. BS::synced_stream class BS::synced_stream is a utility class which can be used to synchronize printing to an output stream by different threads. It has the following member functions ( T is a template parameter pack):
synced_stream(std::ostream& stream = std::cout) : Construct a new synced stream which prints to the given output stream.void print(T&&... items) : 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.void println(T&&... items) : Print any number of items into the output stream, followed by a newline character.In addition, the class comes with two stream manipulators, which are meant to help the compiler figure out which template specializations to use with the class:
BS::synced_stream::endl : An explicit cast of std::endl . Prints a newline character to the stream, and then flushes it. Should only be used if flushing is desired, otherwise a newline character should be used instead.BS::synced_stream::flush : An explicit cast of std::flush . Used to flush the stream. BS::timer class BS::timer is a utility class which can be used to measure execution time for benchmarking purposes. It has the following member functions:
timer() : Construct a new timer and immediately start measuring time.void start() : Start (or restart) measuring time. Note that the timer starts ticking as soon as the object is created, so this is only necessary if we want to restart the clock later.void stop() : Stop measuring time and store the elapsed time since the object was constructed or since start() was last called.std::chrono::milliseconds::rep current_ms() : Get the number of milliseconds that have elapsed since the object was constructed or since start() was last called, but keep the timer ticking.std::chrono::milliseconds::rep ms() : Get the number of milliseconds stored when stop() was last called. This library is under continuous and active development. If you encounter any bugs, or if you would like to request any additional features, please feel free to open a new issue on GitHub and I will look into it as soon as I can.
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.
Many GitHub users have helped improve this project, directly or indirectly, via issues, pull requests, comments, and/or personal correspondence. Please see CHANGELOG.md for links to specific issues and pull requests that have been the most helpful. Thank you all for your contribution! :)
If you found this project useful, please consider starring it on GitHub! This allows me to see how many people are using my code, and motivates me to keep working to improve it.
Copyright (c) 2024 Barak Shoshany. Licensed under the MIT license.
If you use this C++ thread pool library in software of any kind, please provide a link to the GitHub repository in the source code and documentation.
If you use this library in published research, please cite it as follows:
You can use the following BibTeX entry:
@article { Shoshany2024_ThreadPool ,
archiveprefix = { arXiv } ,
author = { Barak Shoshany } ,
doi = { 10.1016/j.softx.2024.101687 } ,
eprint = { 2105.00613 } ,
journal = { SoftwareX } ,
pages = { 101687 } ,
title = { {A C++17 Thread Pool for High-Performance Scientific Computing} } ,
url = { https://www.sciencedirect.com/science/article/pii/S235271102400058X } ,
volume = { 26 } ,
year = { 2024 }
} Please note that the papers on SoftwareX and arXiv are not up to date with the latest version of the library. These publications are only intended to facilitate discovery of this library by scientists, and to enable citing it in scientific research. Documentation for the latest version is provided only by the README.md file in the GitHub repository.
Beginner C++ programmers may be interested in my lecture notes for a course taught at McMaster University, which teach modern C and C++ from scratch, including some of the advanced techniques and programming practices used in developing this library.
