![\"BackgroundCpp\"](\"https:\/\/www.modernescpp.com\/wp-content\/uploads\/2016\/06\/BackgroundCpp.jpg\")
<\/h2>\n<\/span>Multithreading in modern C++<\/span><\/h2>\n\n- Overview<\/a><\/li>\n<\/ul>\n
<\/span>C++ memory model<\/span><\/h3>\n\n\n- Overview<\/a><\/li>\n
- Atomic data types\n
\n- std::atomic_flag<\/a><\/li>\n
- std::atomic<\/a><\/li>\n
- std::atomic<\/a><\/li>\n<\/ol>\n<\/li>\n
- Synchronization and ordering constraints<\/a><\/li>\n
- The different C++ memory models\n
\n- Sequential consistency\n
\n- Theory<\/a><\/li>\n
- Practice<\/a><\/li>\n<\/ol>\n<\/li>\n
- Acquire-release semantic\n
\n- Synchronization and ordering constraints<\/a><\/li>\n
- Transitivity<\/a><\/li>\n
- The special case std::memory_order_consume<\/a><\/li>\n
- The typical error<\/a><\/li>\n<\/ol>\n<\/li>\n
- Wait and Sleep<\/a><\/li>\n<\/ol>\n<\/li>\n
- Fences<\/a>\n
\n- Memory barriers<\/a><\/li>\n
- Acquire- and release memory barriers<\/a><\/li>\n<\/ol>\n<\/li>\n
- Algorithms\n
\n- Blocking and non-blocking algorithms<\/a><\/li>\n
- ABA – A is not the same as A<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/ol>\n
<\/span>The threading interface<\/span><\/h3>\n\n- Threads\n
\n- Creation<\/a><\/li>\n
- Lifetime<\/a><\/li>\n
- Arguments<\/a><\/li>\n
- Sharing data<\/a>\n
\n- Risks <\/a><\/li>\n
- Prefer Locks to Mutexes<\/a><\/li>\n
- Reader-Writer locks<\/a><\/li>\n
- Safe initialization<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/li>\n
- Thread-local data<\/a><\/li>\n
- Condition variables<\/a><\/li>\n
- Tasks<\/a>\n
\n- std::async<\/a><\/li>\n
- std::packaged_task<\/a><\/li>\n
- std::promise and std::future<\/a><\/li>\n
- The special futures std::async<\/a><\/li>\n<\/ol>\n<\/li>\n
- Condition variables versus tasks for the synchronization of threads<\/a><\/li>\n
- Challenges\n
\n- Data races versus race conditions<\/a><\/li>\n
- Malicious data races<\/a><\/li>\n
- Blocking and non-blocking algorithms<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Multithreading with C++17 and C++20<\/span><\/h2>\n\n- An overview<\/a>\n
\n- Parallel algorithm of the Standard Template Library<\/a><\/li>\n
- Atomic smart pointers<\/a><\/li>\n
- std::future extensions<\/a><\/li>\n
- Latches and barriers<\/a><\/li>\n
- Coroutines<\/a><\/li>\n
- Transactional memory<\/a><\/li>\n
- Task blocks<\/a><\/li>\n
- Executors<\/a><\/li>\n
- Unified Futures<\/a><\/li>\n
- std:::jthread<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Application of multithreading<\/span><\/h2>\n\n- The time library\n
\n- Overview<\/a><\/li>\n
- Time point<\/a><\/li>\n
- Time duration<\/a><\/li>\n
- Clock<\/a><\/li>\n
- Wait and Sleep<\/a><\/li>\n<\/ol>\n<\/li>\n
- Ongoing optimization with CppMem\n
\n- Context of ongoing optimization<\/a><\/li>\n
- Introduction to CppMem<\/a><\/li>\n
- Non-atomic variables<\/a><\/li>\n
- Locks<\/a><\/li>\n
- Sequential consistency<\/a><\/li>\n
- Acquire-release semantic<\/a><\/li>\n
- Acquire-release semantic with a data race<\/a><\/li>\n
- Relaxed semantic<\/a><\/li>\n
- volatile<\/a><\/li>\n<\/ol>\n<\/li>\n
- Singleton\n
\n- Thread safe initialization of a singleton<\/a><\/li>\n<\/ol>\n<\/li>\n
- Calculating the sum of a std::vector<\/span>\n
\n- Single threaded<\/a><\/li>\n
- Multithreaded with shared variables<\/a><\/li>\n
- Multithreaded with minimal synchronization<\/a><\/li>\n
- My conclusion<\/a><\/li>\n<\/ol>\n<\/li>\n
- The Dining Philosophers Problem\n
\n- Dining Philosophers Problem I<\/a><\/li>\n
- Dining Philosophers Problem II<\/a><\/li>\n
- Dining Philosophers Problem III<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Embedded programming with C++<\/span><\/h2>\n\n- Requirements for embedded programming<\/a><\/li>\n
- Myths <\/a>and facts<\/a><\/li>\n<\/ul>\n
<\/span>High safety requirements<\/a><\/span><\/h3>\n\n- C++11\n
\n- Automatic type deduction with <\/a>auto<\/a><\/span><\/li>\n
- Prevent narrowing with {} initialization<\/a><\/li>\n
- Guarantees at compile time with static_assert<\/span> and the type-traits library\n
\n- static_assert<\/a><\/li>\n
- Continuously improvement- An introductory example to the type-traits library<\/a><\/li>\n
- Check types<\/a><\/li>\n
- Compare and modify types<\/a><\/li>\n<\/ol>\n<\/li>\n
- User-defined literals\n
\n- Type safe calculation with user-defined literals<\/a><\/li>\n
- Raw and cooked<\/a><\/li>\n<\/ol>\n<\/li>\n
- Strongly-Typed Enumerations<\/a><\/li>\n
- override<\/span> and <\/a>final<\/a><\/span><\/li>\n
- The null pointer constant <\/a>nullptr<\/a><\/span><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Performance matters<\/span><\/h3>\n
<\/span>C++ memory model<\/span><\/h3>\n\n\n- Overview<\/a><\/li>\n
- Atomic data types\n
\n- std::atomic_flag<\/a><\/li>\n
- std::atomic<\/a><\/li>\n
- std::atomic<\/a><\/li>\n<\/ol>\n<\/li>\n
- Synchronization and ordering constraints<\/a><\/li>\n
- The different C++ memory models\n
\n- Sequential consistency\n
\n- Theory<\/a><\/li>\n
- Practice<\/a><\/li>\n<\/ol>\n<\/li>\n
- Acquire-release semantic\n
\n- Synchronization and ordering constraints<\/a><\/li>\n
- Transitivity<\/a><\/li>\n
- The special case std::memory_order_consume<\/a><\/li>\n
- The typical error<\/a><\/li>\n<\/ol>\n<\/li>\n
- Wait and Sleep<\/a><\/li>\n<\/ol>\n<\/li>\n
- Fences<\/a>\n
\n- Memory barriers<\/a><\/li>\n
- Acquire- and release memory barriers<\/a><\/li>\n<\/ol>\n<\/li>\n
- Algorithms\n
\n- Blocking and non-blocking algorithms<\/a><\/li>\n
- ABA – A is not the same as A<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/ol>\n
<\/span>The threading interface<\/span><\/h3>\n\n- Threads\n
\n- Creation<\/a><\/li>\n
- Lifetime<\/a><\/li>\n
- Arguments<\/a><\/li>\n
- Sharing data<\/a>\n
\n- Risks <\/a><\/li>\n
- Prefer Locks to Mutexes<\/a><\/li>\n
- Reader-Writer locks<\/a><\/li>\n
- Safe initialization<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/li>\n
- Thread-local data<\/a><\/li>\n
- Condition variables<\/a><\/li>\n
- Tasks<\/a>\n
\n- std::async<\/a><\/li>\n
- std::packaged_task<\/a><\/li>\n
- std::promise and std::future<\/a><\/li>\n
- The special futures std::async<\/a><\/li>\n<\/ol>\n<\/li>\n
- Condition variables versus tasks for the synchronization of threads<\/a><\/li>\n
- Challenges\n
\n- Data races versus race conditions<\/a><\/li>\n
- Malicious data races<\/a><\/li>\n
- Blocking and non-blocking algorithms<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Multithreading with C++17 and C++20<\/span><\/h2>\n\n- An overview<\/a>\n
\n- Parallel algorithm of the Standard Template Library<\/a><\/li>\n
- Atomic smart pointers<\/a><\/li>\n
- std::future extensions<\/a><\/li>\n
- Latches and barriers<\/a><\/li>\n
- Coroutines<\/a><\/li>\n
- Transactional memory<\/a><\/li>\n
- Task blocks<\/a><\/li>\n
- Executors<\/a><\/li>\n
- Unified Futures<\/a><\/li>\n
- std:::jthread<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Application of multithreading<\/span><\/h2>\n\n- The time library\n
\n- Overview<\/a><\/li>\n
- Time point<\/a><\/li>\n
- Time duration<\/a><\/li>\n
- Clock<\/a><\/li>\n
- Wait and Sleep<\/a><\/li>\n<\/ol>\n<\/li>\n
- Ongoing optimization with CppMem\n
\n- Context of ongoing optimization<\/a><\/li>\n
- Introduction to CppMem<\/a><\/li>\n
- Non-atomic variables<\/a><\/li>\n
- Locks<\/a><\/li>\n
- Sequential consistency<\/a><\/li>\n
- Acquire-release semantic<\/a><\/li>\n
- Acquire-release semantic with a data race<\/a><\/li>\n
- Relaxed semantic<\/a><\/li>\n
- volatile<\/a><\/li>\n<\/ol>\n<\/li>\n
- Singleton\n
\n- Thread safe initialization of a singleton<\/a><\/li>\n<\/ol>\n<\/li>\n
- Calculating the sum of a std::vector<\/span>\n
\n- Single threaded<\/a><\/li>\n
- Multithreaded with shared variables<\/a><\/li>\n
- Multithreaded with minimal synchronization<\/a><\/li>\n
- My conclusion<\/a><\/li>\n<\/ol>\n<\/li>\n
- The Dining Philosophers Problem\n
\n- Dining Philosophers Problem I<\/a><\/li>\n
- Dining Philosophers Problem II<\/a><\/li>\n
- Dining Philosophers Problem III<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Embedded programming with C++<\/span><\/h2>\n\n- Requirements for embedded programming<\/a><\/li>\n
- Myths <\/a>and facts<\/a><\/li>\n<\/ul>\n
<\/span>High safety requirements<\/a><\/span><\/h3>\n\n- C++11\n
\n- Automatic type deduction with <\/a>auto<\/a><\/span><\/li>\n
- Prevent narrowing with {} initialization<\/a><\/li>\n
- Guarantees at compile time with static_assert<\/span> and the type-traits library\n
\n- static_assert<\/a><\/li>\n
- Continuously improvement- An introductory example to the type-traits library<\/a><\/li>\n
- Check types<\/a><\/li>\n
- Compare and modify types<\/a><\/li>\n<\/ol>\n<\/li>\n
- User-defined literals\n
\n- Type safe calculation with user-defined literals<\/a><\/li>\n
- Raw and cooked<\/a><\/li>\n<\/ol>\n<\/li>\n
- Strongly-Typed Enumerations<\/a><\/li>\n
- override<\/span> and <\/a>final<\/a><\/span><\/li>\n
- The null pointer constant <\/a>nullptr<\/a><\/span><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Performance matters<\/span><\/h3>\n
- \n
- Overview<\/a><\/li>\n
- Atomic data types\n
- \n
- std::atomic_flag<\/a><\/li>\n
- std::atomic<\/a><\/li>\n
- std::atomic<\/a><\/li>\n<\/ol>\n<\/li>\n
- Synchronization and ordering constraints<\/a><\/li>\n
- The different C++ memory models\n
- \n
- Sequential consistency\n
- \n
- Theory<\/a><\/li>\n
- Practice<\/a><\/li>\n<\/ol>\n<\/li>\n
- Acquire-release semantic\n
- \n
- Synchronization and ordering constraints<\/a><\/li>\n
- Transitivity<\/a><\/li>\n
- The special case std::memory_order_consume<\/a><\/li>\n
- The typical error<\/a><\/li>\n<\/ol>\n<\/li>\n
- Wait and Sleep<\/a><\/li>\n<\/ol>\n<\/li>\n
- Fences<\/a>\n
- \n
- Memory barriers<\/a><\/li>\n
- Acquire- and release memory barriers<\/a><\/li>\n<\/ol>\n<\/li>\n
- Algorithms\n
- \n
- Blocking and non-blocking algorithms<\/a><\/li>\n
- ABA – A is not the same as A<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/ol>\n
<\/span>The threading interface<\/span><\/h3>\n
- \n
- Threads\n
- \n
- Creation<\/a><\/li>\n
- Lifetime<\/a><\/li>\n
- Arguments<\/a><\/li>\n
- Sharing data<\/a>\n
- \n
- Risks <\/a><\/li>\n
- Prefer Locks to Mutexes<\/a><\/li>\n
- Reader-Writer locks<\/a><\/li>\n
- Safe initialization<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/li>\n
- Thread-local data<\/a><\/li>\n
- Condition variables<\/a><\/li>\n
- Tasks<\/a>\n
- \n
- std::async<\/a><\/li>\n
- std::packaged_task<\/a><\/li>\n
- std::promise and std::future<\/a><\/li>\n
- The special futures std::async<\/a><\/li>\n<\/ol>\n<\/li>\n
- Condition variables versus tasks for the synchronization of threads<\/a><\/li>\n
- Challenges\n
- \n
- Data races versus race conditions<\/a><\/li>\n
- Malicious data races<\/a><\/li>\n
- Blocking and non-blocking algorithms<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Multithreading with C++17 and C++20<\/span><\/h2>\n
- \n
- An overview<\/a>\n
- \n
- Parallel algorithm of the Standard Template Library<\/a><\/li>\n
- Atomic smart pointers<\/a><\/li>\n
- std::future extensions<\/a><\/li>\n
- Latches and barriers<\/a><\/li>\n
- Coroutines<\/a><\/li>\n
- Transactional memory<\/a><\/li>\n
- Task blocks<\/a><\/li>\n
- Executors<\/a><\/li>\n
- Unified Futures<\/a><\/li>\n
- std:::jthread<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Application of multithreading<\/span><\/h2>\n
- \n
- The time library\n
- \n
- Overview<\/a><\/li>\n
- Time point<\/a><\/li>\n
- Time duration<\/a><\/li>\n
- Clock<\/a><\/li>\n
- Wait and Sleep<\/a><\/li>\n<\/ol>\n<\/li>\n
- Ongoing optimization with CppMem\n
- \n
- Context of ongoing optimization<\/a><\/li>\n
- Introduction to CppMem<\/a><\/li>\n
- Non-atomic variables<\/a><\/li>\n
- Locks<\/a><\/li>\n
- Sequential consistency<\/a><\/li>\n
- Acquire-release semantic<\/a><\/li>\n
- Acquire-release semantic with a data race<\/a><\/li>\n
- Relaxed semantic<\/a><\/li>\n
- volatile<\/a><\/li>\n<\/ol>\n<\/li>\n
- Singleton\n
- \n
- Thread safe initialization of a singleton<\/a><\/li>\n<\/ol>\n<\/li>\n
- Calculating the sum of a std::vector<\/span>\n
- \n
- Single threaded<\/a><\/li>\n
- Multithreaded with shared variables<\/a><\/li>\n
- Multithreaded with minimal synchronization<\/a><\/li>\n
- My conclusion<\/a><\/li>\n<\/ol>\n<\/li>\n
- The Dining Philosophers Problem\n
- \n
- Dining Philosophers Problem I<\/a><\/li>\n
- Dining Philosophers Problem II<\/a><\/li>\n
- Dining Philosophers Problem III<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Embedded programming with C++<\/span><\/h2>\n
- \n
- Requirements for embedded programming<\/a><\/li>\n
- Myths <\/a>and facts<\/a><\/li>\n<\/ul>\n
<\/span>High safety requirements<\/a><\/span><\/h3>\n
- \n
- C++11\n
- \n
- Automatic type deduction with <\/a>auto<\/a><\/span><\/li>\n
- Prevent narrowing with {} initialization<\/a><\/li>\n
- Guarantees at compile time with static_assert<\/span> and the type-traits library\n
- \n
- static_assert<\/a><\/li>\n
- Continuously improvement- An introductory example to the type-traits library<\/a><\/li>\n
- Check types<\/a><\/li>\n
- Compare and modify types<\/a><\/li>\n<\/ol>\n<\/li>\n
- User-defined literals\n
- \n
- Type safe calculation with user-defined literals<\/a><\/li>\n
- Raw and cooked<\/a><\/li>\n<\/ol>\n<\/li>\n
- Strongly-Typed Enumerations<\/a><\/li>\n
- override<\/span> and <\/a>final<\/a><\/span><\/li>\n
- The null pointer constant <\/a>nullptr<\/a><\/span><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n
<\/span>Performance matters<\/span><\/h3>\n
- Raw and cooked<\/a><\/li>\n<\/ol>\n<\/li>\n
- Continuously improvement- An introductory example to the type-traits library<\/a><\/li>\n
- Prevent narrowing with {} initialization<\/a><\/li>\n
- Automatic type deduction with <\/a>auto<\/a><\/span><\/li>\n
- Myths <\/a>and facts<\/a><\/li>\n<\/ul>\n
- Dining Philosophers Problem II<\/a><\/li>\n
- Multithreaded with shared variables<\/a><\/li>\n
- Calculating the sum of a std::vector<\/span>\n
- Introduction to CppMem<\/a><\/li>\n
- Time point<\/a><\/li>\n
- Overview<\/a><\/li>\n
- Atomic smart pointers<\/a><\/li>\n
- Parallel algorithm of the Standard Template Library<\/a><\/li>\n
- Malicious data races<\/a><\/li>\n
- std::packaged_task<\/a><\/li>\n
- Prefer Locks to Mutexes<\/a><\/li>\n
- Lifetime<\/a><\/li>\n
- Creation<\/a><\/li>\n
- ABA – A is not the same as A<\/a><\/li>\n<\/ol>\n<\/li>\n<\/ol>\n<\/ol>\n
- Acquire- and release memory barriers<\/a><\/li>\n<\/ol>\n<\/li>\n
- Transitivity<\/a><\/li>\n
- Practice<\/a><\/li>\n<\/ol>\n<\/li>\n
- Theory<\/a><\/li>\n
- std::atomic<\/a><\/li>\n
- Atomic data types\n