• ↑↓ pour naviguer
  • ↵ pour ouvrir
  • ⟶ pour sélectionner
  • ⌘ ⌥ ↵ pour ouvrir dans un panneau
  • ←→ pour naviguer
  • esc pour rejeter
⌘ '
raccourcis clavier
semantic not available yet

thrashing

The Problem When the system is overcommitted (more process demand than RAM), it spends more time swapping pages in/out than doing useful work.

  • published

  • modified

  • length

    1 min read (189 words)

  • source

    llms.txt

The Problem

When the system is overcommitted (more process demand than RAM), it spends more time swapping pages in/out than doing useful work.

Core Idea

Thrashing is a state where the OS spends most of its time handling page faults and swapping, leaving little CPU time for actual process execution.

How It Works

  1. System runs many processes, total memory demand > RAM
  2. Page fault rate increases dramatically
  3. OS responds by swapping pages out to disk
  4. Pages immediately needed again → more page faults
  5. CPU utilization drops (most time spent in page fault handling)
thrashing Normal Normal page faults: few Overload Overload memory demand > RAM Normal->Overload Thrash Thrashing CPU: 10% thrash: 90% Overload->Thrash

Key Properties

  • Caused by overcommitment (too many active processes)
  • CPU utilization drops to near zero
  • System becomes unresponsive
  • Fix: reduce degree of multiprogramming (fewer processes)

Connections

Edge Cases & Gotchas

  • Can cascade: thrashing in one process causes others to thrash too
  • Monitor page fault rate to detect early
  • Fix: suspend/kill processes, or add more RAM

you might also like

  • Binary Search Tree
  • Index for backend
  • HTTP Versions
  • Nginx
  • django-testing-framework
  • collision-resolution
  • Vim Repetition
  • exposed-terminal-problem
  • operator-precedence-parser
    • demand-paging
    The Problem Loading an entire process into RAM at startup is wasteful — many pages may never be used.
    page-fault
    The Problem In paged virtual memory, a process may reference a page that isn’t currently in RAM.
    swap-space
    The Problem RAM is limited and expensive. When all RAM is full, the OS needs somewhere to put inactive processes so active ones can continue running.
    virtual-memory
    The Problem Programs need more memory than physically available, and we want isolation between processes (one process shouldn’t access another’s memory).