From Tal Weiss at Takipi Blog:
The problem with synchronized blocks is that they’re all or nothing – you can’t have more than one thread inside a critical section. This is especially a bummer in consumer/producer scenarios, where some threads are trying to edit some data exclusively, while others are only trying to read it and are fine with sharing access.
ReadWriteLocks were meant to be the perfect solution for this. You can specify which threads block everyone else (writers), and which ones play well with others for consuming content (readers). A happy ending? Afraid not.
Unlike synchronized blocks, RW locks are not built-in to the JVM and have the same capabilities as mere mortal code. Still, to implement a locking idiom you need to instruct the CPU to perform specific operations atomically, or in specific order, to avoid race conditions. This is traditionally done through the magical portal-hole into the JVM – the unsafe class. RW Locks use Compare-And-Swap (CAS) operations to set values directly into memory as part of their thread queuing algorithm
Even so, RWLocks are just not fast enough, and at times prove to be really darn slow, to the point of not being worth bothering with. However help is on the way, with the good folks at the JDK not giving up, and are now back with the new StampedLock. This RW lock employs a new set of algorithms and memory fencing features added to the Java 8 JDK to help make this lock faster and more robust.
Does it deliver on its promise? Let’s see...