/* ThreadWrapper(Test) - starting threads and passing context Copyright (C) 2008, Hermann Vosseler   **Lumiera** 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 2 of the License, or (at your   option) any later version. See the file COPYING for further details. * *****************************************************************/ /** @file thread-wrapper-test.cpp ** unit test \ref ThreadWrapper_test */ #include "lib/test/run.hpp" #include "lib/thread.hpp" #include "lib/iter-explorer.hpp" #include "lib/scoped-collection.hpp" #include "lib/test/microbenchmark.hpp" #include #include using test::Test; using lib::explore; using std::atomic_uint; using std::this_thread::yield; using std::this_thread::sleep_for; using std::chrono::microseconds; namespace lib { namespace test{ namespace { // test parameters const uint NUM_THREADS = 200; const uint REPETITIONS = 10; } /*******************************************************************//** * @test use the lib::Thread wrapper for simplified definition of the * thread-function, argument binding and starting of threads. * @see thread.hpp * @see ThreadWrapperJoin_test * @see SyncLocking_test */ class ThreadWrapper_test : public Test { virtual void run (Arg) { seedRand(); demonstrateSimpleUsage(); verifyConcurrentExecution(); } /** @test demonstrate simple usage of the thread-wrapper a λ-binding */ void demonstrateSimpleUsage() { atomic_uint i{0}; Thread thread("counter", [&]{ ++i; }); // bind a λ and launch thread while (thread) yield(); // ensure thread has finished and detached CHECK (i == 1); // verify the effect has taken place } /** * @test verify the thread function is actually performed concurrently * - use a derived Thread object, also holding a local data field * - the thread function sleeps, and then stores the sum of two numbers * - demonstrate that each instance can have a different argument binding * - verify each thread function has actually been invoked once per thread, * by comparing a local sum with values collected from the Thread objects, * - moreover measure the overall time required for launching the threads * and then waiting for all threads to have terminated and detached; * this time must be _shorter_ than all the _average_ sleep times * compounded (as if the function was invoked sequentially). */ void verifyConcurrentExecution() { struct TestThread : Thread { using Thread::Thread; uint local{0}; void doIt (uint a, uint b) ///< the actual operation running in a separate thread { uint sum = a + b; sleep_for (microseconds{sum}); // Note: explicit random delay before local store local = sum; } }; // prepare Storage for these objects (not created yet) lib::ScopedCollection threads{NUM_THREADS}; size_t checkSum = 0; size_t globalSum = 0; auto launchThreads = [&] { for (uint i=1; i<=NUM_THREADS; ++i) { uint x = rani(1000); globalSum += (i + x); threads.emplace (&TestThread::doIt, i, x); } // Note: bind to member function, copying arguments while (explore(threads).has_any()) yield(); // wait for all threads to have detached for (auto& t : threads) { CHECK (0 < t.local); checkSum += t.local; } }; double runTime = benchmarkTime (launchThreads, REPETITIONS); CHECK (checkSum == globalSum); // sum of precomputed random numbers matches sum from threads CHECK (runTime < NUM_THREADS * 1000/2); // random sleep time should be > 500ms on average } }; /** Register this test class... */ LAUNCHER (ThreadWrapper_test, "function common"); }} // namespace lib::test