Category: Concurrency

Let’s assume you want to implement a function to update the maximum value or a variable. Issue: To prevent a

Advantages … of a singleton is that you don’t have to pass it into every class you want to log

std::atomic Intel / AMD ARM (with LSE) fetch_add() LOCK XADD1,2,3,4,5 LDADD1, LDADDA2, LDADDL3, LDADDAL4,5 fetch_sub() fetch_and() LOCK AND1,2,3,4,5 LDCLR1, LDCLRA2,

With C++20 you can use std::jthread (“Cooperatively Interruptible Joining Thread”), which does the stop request and the joining in its

Code $ g++ main.cpp -ltbb -lpthread Future Switch to a thread pool to execute the different calculations in parallel.

Overview Memory order Description memory_order_relaxed Unsafest, No ordering is guaranteed. memory_order_acquire A load operation, No reads or writes in the

oneTBB The library abstracts all the thread handling from you, like the task scheduler and its thread pool. Features (selection)

Mutexes (Selection) … synchronizes the access on critical sections std::mutex Provides exclusive access (can only be pulled by one thread

Active waiting Don’t use this approach, because it will unnecessarily consume CPU time. The sleep reduces the CPU consumption, but

The std::mutex is a non-recursive mutex and therefore can’t be locked again from the same thread before getting unlocked. This