Diploma Notes

Operating Systems - Unit 3

3 min read

Academy - OSY 3.1: CPU and I/O Burst Cycles

1. The Core Concept: Process Execution Rhythm

The execution of any process is not a continuous, unbroken use of the CPU. Instead, it is a sequence of alternating cycles:

  1. CPU Burst: A period where the process is actively using the CPU to perform computations (e.g., adding numbers, running loops, processing data).
  2. I/O Burst: A period where the process is waiting for an Input/Output operation to complete (e.g., reading a file from the disk, waiting for a network packet, getting user input).

This continuous switching between CPU work and I/O waiting is known as the CPU–I/O Burst Cycle.

2. Visualizing the Cycle

Imagine our autocleanse.py script running. Its lifecycle looks like this:

[Load Program] -> CPU Burst [Read CSV from Disk] -> I/O Burst (Waiting) [Calculate Null Percentages] -> CPU Burst [Calculate Median] -> CPU Burst [Print to Screen] -> I/O Burst (Waiting) [... and so on ...] [End Program] -> CPU Burst

3. Process Characterization (The “Why It Matters”)

The relative length and frequency of these bursts determine the “personality” of a process. The OS Scheduler needs to know this to manage the system efficiently.

  • I/O-Bound Process:

    • Description: A process that spends most of its time in I/O bursts, with very short CPU bursts.
    • Examples: A web server waiting for requests, a text editor waiting for user input, a script that mainly reads/writes files.
    • Scheduling Goal: The scheduler should run this process frequently for short periods so its I/O requests can be started quickly.

  • CPU-Bound Process:

    • Description: A process that spends most of its time in long CPU bursts, with very infrequent I/O waits.
    • Examples: Scientific simulations, complex mathematical calculations, video rendering.
    • Scheduling Goal: The scheduler should give this process longer, uninterrupted time slots on the CPU to let it finish its work.

4. Connection to Our Experiment

Connection to Practice

We practically demonstrated this concept in our AutoCleanSE project. By running the script on a large, 5-million-row file, we were able to measure the time spent in CPU bursts versus I/O bursts and prove that our script was primarily CPU-bound.

See the full experiment here: Day 3 : Performance Testing at Scale

Our time command output gave us the evidence to characterize our script:

  • real: 16.564s
  • user (CPU Burst Time): 10.51s
  • sys (CPU Burst Time for OS tasks): 3.32s
  • Waiting Time (I/O Burst Time): 16.564 - (10.51 + 3.32) = ~2.7s

The evidence shows that our script spent much more time in CPU Bursts (13.83s total) than in I/O Bursts (~2.7s). Therefore, we can definitively characterize our autocleanse.py script, when running on a large file, as being primarily CPU-bound. This is a critical insight for any future optimization efforts.

Graph View

Backlinks