So, I have been reading a lot of books lately. Java Concurrency in Practice was probably one of the best book written for concurrent programming in Java. Though this book is written prior to Java 7, it is still valid and did an excellent job explaining all the fundamentals.

While reading, I came across a paragraph entitled "Nonatomic 64-bit operations." I actually never knew that in Java 64-bit, long and double values were treated as two 32-bit values.  That means, a 64-bit write operation is basically performed as two separate 32-bit operations. This behavior can result in indeterminate values being read in code and that lacks atomicity.

According to the Java Language Specification (JSL-17.7)

For the purposes of the Java programming language memory model, a single write to a non-volatile long or double value is treated as two separate writes: one to each 32-bit half. This can result in a situation where a thread sees the first 32 bits of a 64-bit value from one write, and the second 32 bits from another write.

Writes and reads of volatile long and double values are always atomic.
Writes to and reads of references are always atomic, regardless of whether they are implemented as 32-bit or 64-bit values.
Some implementations may find it convenient to divide a single write action on a 64-bit long or double value into two write actions on adjacent 32-bit values. For efficiency’s sake, this behavior is implementation-specific; an implementation of the Java Virtual Machine is free to perform writes to long and double values atomically or in two parts.
Implementations of the Java Virtual Machine are encouraged to avoid splitting 64-bit values where possible. Programmers are encouraged to declare shared 64-bit values as volatile or synchronize their programs correctly to avoid possible complications.

So point is, in Java, long and double aren’t thread safe. When multiple threads are going to access a long or a double value without synchronization, it can cause problems. To ensure atomic/thread safety, it is essential to use volatile to ensure changes made by one thread are visible to other threads. That’s how we can ensure the read/write is atomic and thread safe.

However, if all reads and writes of 64-bit long and double values occur within a synchronized block, the atomicity of the read/write is guaranteed.