13.3 Imagine you are writing the code to manage a hash table that will be shared among several concurrent threads. Assume that operations on the table need to be atomic. You could use a single mutual...

Programming Language Pragmatics, 4th Edition

Extracted text: 13.3 Imagine you are writing the code to manage a hash table that will be shared among several concurrent threads. Assume that operations on the table need to be atomic. You could use a single mutual exclusion lock to protect the entire table, or you could devise a scheme with one lock per hash-table bucket. Which approach is likely to work better, under what circumstances? Why?