For example, we might be unhappy if we deposit our paycheck on Monday, but the bank bounces our rent check the following Friday because it reordered our deposit after your withdrawal. (查看原文)
Now consider the behavior of the TTASLock algorithm while the lock is held by a thread A. The first time thread B reads the lock it takes a cache miss, forcing B to block while the value is loaded into B's cache. As long as A holds the lock, B repeatly reread the value, but hits in the cache every time. (查看原文)
This notion of local spinning, where threads repeatly reread cached values instead of repeatedly using the bus, is an important principle critical to the design of efficeient spin locks.
Here is a key observation: if some other thread aquires the lock between hte first and second step, then, most likely, there is high conteniton for that lock.
the larger the number of unsuccessful tries, the higher the likely contention, and the longer the thread should back off .
backing off is common to several locking algorithms.
the BackoffLock is easy to implement, and typically performs significantly better than TASLock on many architectures. (查看原文)
