
Baeldung Pro comes with both absolutely No-Ads as well as finally with Dark Mode, for a clean learning experience:
Once the early-adopter seats are all used, the price will go up and stay at $33/year.
Last updated: December 24, 2024
The compiler, runtime, or processors may apply all sorts of optimizations. Although these optimizations are usually beneficial, typically without necessary synchronizations, they can cause subtle issues in the form of unexpected results.
Caching and reordering are optimizations that may surprise us in concurrent contexts when we don’t synchronize code correctly. Java and the JVM provide many ways to control memory order, and a volatile field is one of them.
This tutorial focuses on Java’s foundational but often misunderstood concept, the volatile field. First, we’ll start with some background about how the underlying computer architecture works, and then we’ll get familiar with memory order in Java. Further, we’ll understand the challenges of concurrency in multiprocessor shared architecture and how volatile fields help fix them.
Processors are responsible for executing program instructions. Therefore, they must retrieve the program instructions and required data from RAM.
As CPUs can carry many instructions per second, fetching from RAM isn’t ideal for them. To improve this situation, processors use tricks like Out of Order Execution, Branch Prediction, Speculative Execution, and Caching.
This is where the following memory hierarchy comes into play:
As different cores execute more instructions and manipulate more data, they fill their caches with more relevant data and instructions. This will improve the overall performance at the expense of introducing cache coherence challenges.
We should think twice about what happens when one thread updates a cached value.
To expand more on cache coherence, we’ll borrow an example from the book Java Concurrency in Practice:
public class TaskRunner {
private static int number;
private static boolean ready;
private static class Reader extends Thread {
@Override
public void run() {
while (!ready) {
Thread.yield();
}
System.out.println(number);
}
}
public static void main(String[] args) {
new Reader().start();
number = 42;
ready = true;
}
}
The TaskRunner class maintains two simple variables. Its main method creates another thread that spins on the ready variable as long as it’s false. When the variable becomes true, the thread prints the number variable.
Many may expect this program to print 42 after a short delay; however, the delay may be much longer. It may even hang forever or print zero.
The cause of these anomalies is the lack of proper memory visibility and reordering. Let’s evaluate them in more detail.
This simple example has two application threads: the main and the reader threads. Let’s imagine a scenario in which the OS schedules those threads on two different CPU cores, where the:
Most modern processors write requests that won’t be applied right after they’re issued. Processors tend to queue those writes in a special write buffer. After a while, they’ll apply those writes to the main memory all at once.
With all that being said, when the main thread updates the number and ready variables, there’s no guarantee about what the reader thread may see. In other words, the reader thread may immediately see the updated value, with some delay, or never at all.
This memory visibility may cause liveness issues in programs relying on visibility.
To make matters even worse, the reader thread may see those writes in an order other than the actual program order. For instance, since we first update the number variable:
public static void main(String[] args) {
new Reader().start();
number = 42;
ready = true;
}
We may expect the reader thread to print 42. But it’s actually possible to see zero as the printed value.
Reordering is an optimization technique for performance improvements. Interestingly, different components may apply this optimization:
We can use volatile to tackle the issues with Cache Coherence.
To ensure that updates to variables propagate predictably to other threads, we should apply the volatile modifier to those variables.
This way, we can communicate with runtime and processor to avoid reordering any instruction involving the volatile variable. Also, processors understand that they should immediately flush any updates to these variables.
public class TaskRunner {
private volatile static int number;
private volatile static boolean ready;
// same as before
}
This way, we communicate with runtime and processor to avoid reordering any instruction involving the volatile variable. Also, processors understand that they should immediately flush any updates to these variables.
For multithreaded applications, we need to ensure a couple of rules for consistent behaviour:
The synchronized methods and blocks provide both of the above properties at the cost of application performance.
The volatile field is quite a useful mechanism because it can help ensure the visibility aspect of the data change without providing mutual exclusion. Thus, it’s useful where we’re ok with multiple threads executing a block of code in parallel, but we need to ensure the visibility property.
A Java compiler is allowed to reorder program text instructions if, in a given thread, it does not affect the execution of that thread in isolation. A shared variable that includes the volatile modifier guarantees that, for each thread, the runtime and processor carry out the instructions related to the shared variable in the same order as they appear in the program text without applying any optimizations that may reorder the instructions. A shared variable that includes the volatile modifier guarantees that all threads see a consistent value for the shared variable. Any update to a volatile field updates the shared value of the field immediately. In other words, a different thread cannot get an inconsistent value of the shared variable after its value is updated.
It’s important to understand that volatile guarantees a consistent value of a shared variable; however, the absence of the volatile modifier doesn’t mean or guarantee that multiple threads always get an inconsistent value of the shared variable. In the absence of the volatile modifier, other threads may occasionally get an inconsistent value of the shared variable. Therefore, let’s not expect that by removing the volatile modifier in the example application TaskRunner, the application starts to print inconsistent values for a shared variable; however, it could happen.
The memory visibility effects of volatile variables extend beyond the volatile variables themselves.
To make matters more concrete, suppose thread A writes to a volatile variable, and then thread B reads the same volatile variable. In such cases, the values that were visible to A before writing the volatile variable will be visible to B after reading the volatile variable:
Technically, any write to a volatile field happens before every subsequent read of the same field. This is the volatile variable rule of the Java Memory Model (JMM).
Because of the strength of the happens-before memory ordering, sometimes we can piggyback on the visibility properties of another volatile variable. For instance, in our particular example, we just need to mark the ready variable as volatile:
public class TaskRunner {
private static int number; // not volatile
private volatile static boolean ready;
// same as before
}
Anything prior to writing true to the ready variable is visible to anything after reading the ready variable. Therefore, the number variable piggybacks on the memory visibility enforced by the ready variable. Simply put, even though it’s not a volatile variable, it’s exhibiting a volatile behaviour.
Using these semantics, we can define only a few variables in our class as volatile and optimize the visibility guarantee.
In this article, we explored the volatile field keyword, its capabilities, and the improvements made to it, starting with Java 5.