// -*- C++ -*- // Copyright (C) 2007, 2008, 2009 Free Software Foundation, Inc. // // This file is part of the GNU ISO C++ Library. This library is free // software; you can redistribute it and/or modify it under the terms // of the GNU General Public License as published by the Free Software // Foundation; either version 3, or (at your option) any later // version. // This library is distributed in the hope that it will be useful, but // WITHOUT ANY WARRANTY; without even the implied warranty of // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU // General Public License for more details. // Under Section 7 of GPL version 3, you are granted additional // permissions described in the GCC Runtime Library Exception, version // 3.1, as published by the Free Software Foundation. // You should have received a copy of the GNU General Public License and // a copy of the GCC Runtime Library Exception along with this program; // see the files COPYING3 and COPYING.RUNTIME respectively. If not, see // . /** @file parallel/workstealing.h * @brief Parallelization of embarrassingly parallel execution by * means of work-stealing. * * Work stealing is described in * * R. D. Blumofe and C. E. Leiserson. * Scheduling multithreaded computations by work stealing. * Journal of the ACM, 46(5):720–748, 1999. * * This file is a GNU parallel extension to the Standard C++ Library. */ // Written by Felix Putze. #ifndef _GLIBCXX_PARALLEL_WORKSTEALING_H #define _GLIBCXX_PARALLEL_WORKSTEALING_H 1 #include #include #include namespace __gnu_parallel { #define _GLIBCXX_JOB_VOLATILE volatile /** @brief One job for a certain thread. */ template struct Job { typedef _DifferenceTp difference_type; /** @brief First element. * * Changed by owning and stealing thread. By stealing thread, * always incremented. */ _GLIBCXX_JOB_VOLATILE difference_type first; /** @brief Last element. * * Changed by owning thread only. */ _GLIBCXX_JOB_VOLATILE difference_type last; /** @brief Number of elements, i. e. @c last-first+1. * * Changed by owning thread only. */ _GLIBCXX_JOB_VOLATILE difference_type load; }; /** @brief Work stealing algorithm for random access iterators. * * Uses O(1) additional memory. Synchronization at job lists is * done with atomic operations. * @param begin Begin iterator of element sequence. * @param end End iterator of element sequence. * @param op User-supplied functor (comparator, predicate, adding * functor, ...). * @param f Functor to "process" an element with op (depends on * desired functionality, e. g. for std::for_each(), ...). * @param r Functor to "add" a single result to the already * processed elements (depends on functionality). * @param base Base value for reduction. * @param output Pointer to position where final result is written to * @param bound Maximum number of elements processed (e. g. for * std::count_n()). * @return User-supplied functor (that may contain a part of the result). */ template Op for_each_template_random_access_workstealing(RandomAccessIterator begin, RandomAccessIterator end, Op op, Fu& f, Red r, Result base, Result& output, typename std::iterator_traits :: difference_type bound) { _GLIBCXX_CALL(end - begin) typedef std::iterator_traits traits_type; typedef typename traits_type::difference_type difference_type; const _Settings& __s = _Settings::get(); difference_type chunk_size = static_cast(__s.workstealing_chunk_size); // How many jobs? difference_type length = (bound < 0) ? (end - begin) : bound; // To avoid false sharing in a cache line. const int stride = __s.cache_line_size * 10 / sizeof(Job) + 1; // Total number of threads currently working. thread_index_t busy = 0; Job *job; omp_lock_t output_lock; omp_init_lock(&output_lock); // Write base value to output. output = base; // No more threads than jobs, at least one thread. thread_index_t num_threads = __gnu_parallel::max(1, __gnu_parallel::min(length, get_max_threads())); # pragma omp parallel shared(busy) num_threads(num_threads) { # pragma omp single { num_threads = omp_get_num_threads(); // Create job description array. job = new Job[num_threads * stride]; } // Initialization phase. // Flags for every thread if it is doing productive work. bool iam_working = false; // Thread id. thread_index_t iam = omp_get_thread_num(); // This job. Job& my_job = job[iam * stride]; // Random number (for work stealing). thread_index_t victim; // Local value for reduction. Result result = Result(); // Number of elements to steal in one attempt. difference_type steal; // Every thread has its own random number generator // (modulo num_threads). random_number rand_gen(iam, num_threads); // This thread is currently working. # pragma omp atomic ++busy; iam_working = true; // How many jobs per thread? last thread gets the rest. my_job.first = static_cast(iam * (length / num_threads)); my_job.last = (iam == (num_threads - 1)) ? (length - 1) : ((iam + 1) * (length / num_threads) - 1); my_job.load = my_job.last - my_job.first + 1; // Init result with first value (to have a base value for reduction). if (my_job.first <= my_job.last) { // Cannot use volatile variable directly. difference_type my_first = my_job.first; result = f(op, begin + my_first); ++my_job.first; --my_job.load; } RandomAccessIterator current; # pragma omp barrier // Actual work phase // Work on own or stolen start while (busy > 0) { // Work until no productive thread left. # pragma omp flush(busy) // Thread has own work to do while (my_job.first <= my_job.last) { // fetch-and-add call // Reserve current job block (size chunk_size) in my queue. difference_type current_job = fetch_and_add(&(my_job.first), chunk_size); // Update load, to make the three values consistent, // first might have been changed in the meantime my_job.load = my_job.last - my_job.first + 1; for (difference_type job_counter = 0; job_counter < chunk_size && current_job <= my_job.last; ++job_counter) { // Yes: process it! current = begin + current_job; ++current_job; // Do actual work. result = r(result, f(op, current)); } # pragma omp flush(busy) } // After reaching this point, a thread's job list is empty. if (iam_working) { // This thread no longer has work. # pragma omp atomic --busy; iam_working = false; } difference_type supposed_first, supposed_last, supposed_load; do { // Find random nonempty deque (not own), do consistency check. yield(); # pragma omp flush(busy) victim = rand_gen(); supposed_first = job[victim * stride].first; supposed_last = job[victim * stride].last; supposed_load = job[victim * stride].load; } while (busy > 0 && ((supposed_load <= 0) || ((supposed_first + supposed_load - 1) != supposed_last))); if (busy == 0) break; if (supposed_load > 0) { // Has work and work to do. // Number of elements to steal (at least one). steal = (supposed_load < 2) ? 1 : supposed_load / 2; // Push victim's start forward. difference_type stolen_first = fetch_and_add( &(job[victim * stride].first), steal); difference_type stolen_try = stolen_first + steal - difference_type(1); my_job.first = stolen_first; my_job.last = __gnu_parallel::min(stolen_try, supposed_last); my_job.load = my_job.last - my_job.first + 1; // Has potential work again. # pragma omp atomic ++busy; iam_working = true; # pragma omp flush(busy) } # pragma omp flush(busy) } // end while busy > 0 // Add accumulated result to output. omp_set_lock(&output_lock); output = r(output, result); omp_unset_lock(&output_lock); } delete[] job; // Points to last element processed (needed as return value for // some algorithms like transform) f.finish_iterator = begin + length; omp_destroy_lock(&output_lock); return op; } } // end namespace #endif /* _GLIBCXX_PARALLEL_WORKSTEALING_H */