A is the child (created by cilk_spawn), B is the continuation. They sync up at cilk_sync to get the information back to the root. Were it to be executed serially, the result would be the same; the idea is that forking the child will allow A and B to be computed in parallel.
A is the child (created by cilk_spawn), B is the continuation. They sync up at cilk_sync to get the information back to the root. Were it to be executed serially, the result would be the same; the idea is that forking the child will allow A and B to be computed in parallel.