With the acquire-releae semantic, we break the sequential consistency. In the acquire-release semantic the synchronization takes place between atomic operations on the same atomic and not between threads.
With atomic data types you can tailor your program to your needs and therefore optimize it. But now we are in the domain of the multithreading experts.
The easiest way to solve the undefined behaviour in the post Ongoing Optimization: Unsynchronized access is, to use a lock.
I've described my challenge in the last post. Let' start with our process of ongoing optimization. To be sure, I verify my reasoning with CppMem. I once made a big mistake in my presentation at Meeting C++ 2014.
Now it's time to put the theory into practice. The job is quite easy. A small program should undergo an ongoing optimization.
CppMem is an interactive tool for exploring the behaviour of small code snippets of the C++ memory model. It should, no it has to be in the tool box of each programmer, who deals seriously with the memory model.
The relaxed semantic is the end of the Scala. The relaxed semantic is the weakest of all memory models and guarantees only, that the operations on atomic variables are atomic.
Acquire and release fences guarantees similar synchronisation and ordering constraints as atomics with acquire-release semantic. Similar, because the differences are in the details.
The key idea of a std::atomic_thread_fence is, to establish synchronisation and ordering constraints between threads without an atomic operation.
Currently are 154 guests and no members online
Kubik-Rubik Joomla! Extensions