Threads in OS — Process vs Thread, Multithreading | Tutorials Logic

Process vs Thread

A process is an independent program in execution with its own memory space. A thread is the smallest unit of execution within a process. Multiple threads within the same process share the process's resources.

Feature	Process	Thread
Memory	Separate address space	Shared address space (within process)
Resources	Own file handles, I/O, etc.	Shares process resources
Creation overhead	High (fork/exec)	Low (lightweight)
Communication	IPC (pipes, sockets, shared memory)	Direct (shared memory)
Isolation	Crash doesn't affect other processes	Crash can kill entire process
Context switch	Expensive (full address space switch)	Cheaper (same address space)

What each thread has (private): Thread ID, program counter, register set, stack

What threads share (with process): Code section, data section, heap, open files, signals

Benefits of Multithreading

Responsiveness: A thread can continue running while another is blocked (e.g., waiting for I/O). UI remains responsive.
Resource sharing: Threads share memory - no need for expensive IPC mechanisms.
Economy: Creating and context-switching threads is cheaper than processes.
Scalability: Threads can run in parallel on multiple CPU cores (true parallelism).

User Threads vs Kernel Threads

Feature	User-Level Threads	Kernel-Level Threads
Management	Managed by user-space thread library	Managed by OS kernel
Kernel awareness	Kernel sees only one process	Kernel knows about each thread
Context switch	Fast (no kernel mode switch)	Slower (kernel involvement)
Blocking	If one thread blocks, all threads block	Other threads can continue
Parallelism	Cannot run on multiple CPUs simultaneously	Can run on multiple CPUs
Examples	POSIX Pthreads (user-space), Java green threads	Windows threads, Linux pthreads (kernel)

Threading Models

Threading models define how user threads map to kernel threads:

Many-to-One (M:1): Many user threads map to one kernel thread. Fast but no parallelism; one blocking call blocks all threads. (Green threads)
One-to-One (1:1): Each user thread maps to one kernel thread. True parallelism; more overhead. (Linux pthreads, Windows threads)
Many-to-Many (M:N): Many user threads map to many (but fewer) kernel threads. Best of both worlds; complex to implement. (Solaris, older Java)
Two-Level Model: Hybrid of M:N and 1:1 - allows binding specific user threads to kernel threads.

Thread Pool

A thread pool is a collection of pre-created threads waiting for tasks. Instead of creating a new thread for each task (expensive), tasks are submitted to the pool and executed by available threads.

Benefits:

Eliminates thread creation/destruction overhead
Limits the number of concurrent threads (prevents resource exhaustion)
Improves response time (threads are ready immediately)

Thread pool parameters:

Core pool size: Minimum number of threads always kept alive
Maximum pool size: Maximum number of threads allowed
Queue: Holds tasks when all threads are busy
Keep-alive time: How long idle threads above core size are kept before termination

Race Conditions in Multithreading

A race condition occurs when two or more threads access shared data concurrently and the result depends on the order of execution. This leads to unpredictable, incorrect behavior.

Example: Two threads both read a counter (value=5), both increment it, and both write back. Expected result: 7. Actual result: 6 (one increment is lost).

Solutions:

Mutex/Lock: Only one thread can access the critical section at a time
Atomic operations: Hardware-level operations that complete without interruption (e.g., AtomicInteger in Java)
Synchronized methods: Java's synchronized keyword
Immutable objects: Objects that cannot be modified after creation are inherently thread-safe