diff --git a/src/lib/sync-barrier.hpp b/src/lib/sync-barrier.hpp index 4135abb36..9782d6120 100644 --- a/src/lib/sync-barrier.hpp +++ b/src/lib/sync-barrier.hpp @@ -48,7 +48,7 @@ //#include "lib/meta/function.hpp" //#include "lib/result.hpp" -//#include +#include #include @@ -62,20 +62,37 @@ namespace lib { /** - * @todo write type comment + * A one time N-fold mutual synchronisation barrier. + * Calls to #sync() will block until N such calls occurred. + * @note The blocking wait is implemented by a check-and-`yield()` loop, + * increasing load at the OS scheduler, possibly starving the system + * when stretched out over extended time. + * @remark intended use is to allow all participants to catch up and reach + * a well defined point with initialisation or implementation logic. */ class SyncBarrier : util::NonCopyable { + std::atomic_int latch_; + public: + /** @param nFold the number of participants to sync */ explicit - SyncBarrier (size_t nFold =2) - { } + SyncBarrier (uint nFold =2) + : latch_{int(nFold)} + { + REQUIRE (nFold >= 2, "Pointless to sync less than two participants."); + } void sync() { - UNIMPLEMENTED ("count-down latch logic"); + size_t level = latch_.fetch_add(-1, std::memory_order_acq_rel); + if (1 < level) + do std::this_thread::yield(); + while (0 < latch_.load (std::memory_order_relaxed)); + else + latch_.store (0, std::memory_order_relaxed); } }; diff --git a/tests/library/sync-barrier-test.cpp b/tests/library/sync-barrier-test.cpp index bf5fb3538..289e03b5d 100644 --- a/tests/library/sync-barrier-test.cpp +++ b/tests/library/sync-barrier-test.cpp @@ -26,15 +26,10 @@ #include "lib/test/run.hpp" -#include "lib/test/test-helper.hpp" - -//#include "lib/thread.hpp" #include "lib/sync-barrier.hpp" #include "lib/iter-explorer.hpp" #include "lib/util-foreach.hpp" -#include "lib/error.hpp" -//#include #include #include #include @@ -43,9 +38,9 @@ using test::Test; using util::and_all; using lib::explore; -using std::atomic_uint; using std::array; +using std::atomic_uint; using std::this_thread::sleep_for; using namespace std::chrono_literals; @@ -53,7 +48,7 @@ using namespace std::chrono_literals; namespace lib { namespace test { - namespace { + namespace {// Test setup for a concurrent calculation with checksum.... const uint NUM_THREADS = 1024; @@ -61,7 +56,7 @@ namespace test { atomic_uint stage2{0}; atomic_uint finish{0}; - SyncBarrier interThread{NUM_THREADS}; + SyncBarrier interThread{NUM_THREADS }; SyncBarrier afterThread{NUM_THREADS+1}; /** @@ -76,18 +71,18 @@ namespace test { public: TestThread() : thread{[&]() - { //-STAGE-1------------------------------ - localSum = rand() % 1000; // generate local value - stage1.fetch_add (localSum); // book in local value - interThread.sync(); // wait for all other threads to have booked in + { //-STAGE-1------------------------------ + localSum = rand() % 1000; // generate local value + stage1.fetch_add (localSum); // book in local value + interThread.sync(); // wait for all other threads to have booked in - //-STAGE-2------------------------------ - localSum += stage1; // pick up compounded sum from STAGE-1 - localSum += rand() % 1000; // add further local value for STAGE-2 - stage2.fetch_add (localSum); // book in local sum - afterThread.sync(); // wait for other threads and supervisor + //-STAGE-2------------------------------ + uint sync = stage1; // pick up compounded sum from STAGE-1 + localSum += rand() % 1000; // add further local value for STAGE-2 + stage2.fetch_add (localSum+sync); // book in both local values and synced sum + afterThread.sync(); // wait for other threads and supervisor - finish.fetch_add(1); // mark completion of this thread + finish.fetch_add(1); // mark completion of this thread thread::detach(); //////////////////////////////////////////////OOO Wech-oh }} { } @@ -97,7 +92,7 @@ namespace test { }; - /** sum up all `localSum` fields from all TestThread instances in a container */ + /** sum up all `localSum` fields from all TestThread instances in a container */ template uint sumLocals (CON const& threads) @@ -107,12 +102,17 @@ namespace test { .foreach ([&](TestThread const& t){ sum += t.localSum; }); return sum; } - } + }//(End)Test setup + + /*******************************************************************//** * @test verify N-fold synchronisation points by multi-thread loat-test. - * + * - start a _huge number_ of TestThread + * - all those pick up the partial sum from stage1 + * @remark without coordinated synchronisation, some threads would see + * an incomplete sum and thus the stage2 checksum would be lower * @see lib::SyncBarrier * @see steam::control::DispatcherLoop */ @@ -128,7 +128,7 @@ namespace test { CHECK (and_all (threads, [](auto& t){ return t.isRunning(); })); afterThread.sync(); - sleep_for (1ms); // give the threads a chance to terminate + sleep_for (5ms); // give the threads a chance to terminate CHECK (NUM_THREADS == finish); // all threads have passed out.... CHECK (0 < stage1); diff --git a/wiki/thinkPad.ichthyo.mm b/wiki/thinkPad.ichthyo.mm index 9af43b96d..9cc64103d 100644 --- a/wiki/thinkPad.ichthyo.mm +++ b/wiki/thinkPad.ichthyo.mm @@ -79662,8 +79662,7 @@ Date:   Thu Apr 20 18:53:17 2023 +0200
- - + @@ -79699,8 +79698,8 @@ Date:   Thu Apr 20 18:53:17 2023 +0200
 - - + + @@ -79733,9 +79732,9 @@ Date:   Thu Apr 20 18:53:17 2023 +0200
 - + - + @@ -79838,14 +79837,30 @@ Date:   Thu Apr 20 18:53:17 2023 +0200
- - - + + + - + + + + + + + +

+ nicht mutmaßen — messen! +

+ + + + + + +