CS2005 - Lecture 10 - Threads

Concurrency

Managing multiple tasks at the same time, but not necessarily doing them literally at the same instant

Parallelism

Doing multiple things at the exact same time, using multiple CPU cores

Can an application be concurrent but not parallel?

Yes e.g. multitasking on a single-core CPU

What is a thread?

A mini-program running inside a main program. Programs can have multiple threads doing tasks at the same time

Why use multithreading?

Responsiveness: Avoid blocking

Resource Sharing: Threads share memory by default

Economy: Faster to create/switch than processes

Scalability: Utilises multiple cores effectively

What are the states of a Java thread?

NEW: Created but not started

RUNNABLE: Ready or running (scheduler decides)

BLOCKED: Waiting for a monitor lock

WAITING: Indefinite wait

TIMED_WAITING: Waits with timeout

TERMINATED: Completed execution

What is Amdahl's Law?

Speedup <= 1/ S+[(1-S)/N], S=serial portion, N=cores. Limits speedup due to serial code

Name multicore programming challenges

Load balancing, Data dependencies/synchronisation, debugging parallel tasks

Why use multithreading in servers?

Single-threaded: Poor responsiveness (one client at a time)

Process-per-request: Heavy overhead

Multithreaded: Efficiently handle multiple requests (one thread per request)

Creating Threads in Java, Method 1: Extend Thread class

class MyTask extends Thread { public void run() { System.out.println("Task is running!"); } } public class Main { public static void main(String[] args) { MyTask t = new MyTask(); t.start(); } }

Creating Threads in Java, Method 2: Implement Runnable

class MyTask implements Runnable { public void run() { System.out.println("Task is running"); } } public class Main { public static void main(String[] args) { Threads t = new Thread(new MyTask()); t.start(); } }