Thanks for the try gentlemen. :)
I'm trying to understand how using a dispatch barrier on a write op prevents a race condition. It seems to me like converting the queue from concurrent to serial for the write, still means the paused threads can be using a stale value once concurrency is restored. But, heck I'm a chemist. What do I know about threads, processor clock cycles, etc, LOL.
I'm trying to understand how using a dispatch barrier on a write op prevents a race condition. It seems to me like converting the queue from concurrent to serial for the write, still means the paused threads can be using a stale value once concurrency is restored. But, heck I'm a chemist. What do I know about threads, processor clock cycles, etc, LOL.
Comment