Processes vs Threads: Key Differences Every Developer Should Know

This article was originally published on bmf-tech.com.

Overview

Summarizing the differences between processes and threads

What is a Process

• The unit of execution for a program
• Executed on the CPU
• Cannot share resources
• Uses dedicated memory space
  • Holds a virtual address space
  • Switching processes requires mapping between virtual and physical addresses

Data Structures Held by a Process

The data structures held in memory by a process are divided into two segments.

• Text Segment

  • Sequence of instructions for the program (the program itself)
    • Read-only

• Data Segment

  • PDA (Processor Data Area)
    • Contains processor information and data for process management
      • Holds stack pointer, program counter, etc.
        • Stack Pointer
        • Program Counter
  • Data Area
    • Static Area
      • Holds constants and global variables.
      • Heap Area
        • Holds regular variables, etc.
        • The size is unknown at runtime as the process dynamically increases or decreases the area.
  • Stack Area
    • Temporarily holds arguments and local scope data.

What is a Thread

• An execution unit generated from a single process
• Can share resources
• Utilizes shared memory space
  • Holds data within the parent process's memory space.

Differences Between Processes and Threads During Parallel Execution

The differences between processes and threads when executing a program in parallel.

Process

• Can run programs in parallel by forking the parent process to launch multiple child processes.
• When launching a child process, memory space for the child process is allocated.
  • The allocated memory space copies the parent process's data segment and secures a dedicated data segment for the child process.
• Since the program instructions themselves are the same as the parent process, the text segment references the same area as the parent process.
• A child process is still a single process, so it cannot directly access the memory of other processes.

Thread

• Can run programs in parallel by generating threads from a process.
• When generating a thread, the following values are copied into the parent process's virtual address space:
  • Stack Area
  • Stack Pointer
  • Program Counter
• Data other than the above is shared with the parent process.
• A thread only holds data about which part of the program it is executing.
• There is a possibility that a variable a thread intends to use may be modified by another thread.
  • Programs that operate without issues in such situations are called thread-safe.

Performance

Process

• Process switching
  • Requires clearing the cache held by the MMU.
    • TLB flush.

Thread

• Thread switching
  • Only requires switching the stack area, stack pointer, and program counter.

Aside

Stack

• Automatically allocated and freed by the OS or compiler.
• The size is determined during program compilation and linking.

Heap

• The application must handle the allocation when memory is needed and release it when it is no longer needed.
• It is possible to specify the size dynamically when allocating memory.

References

• Differences Between Processes and Threads
• Let's Become Cool Engineers Series - Memory, Processes, and Threads Edition
• Heap and Stack