Welcome to the Parallel and Computing lesson hub! This collection covers three essential paradigms for modern computer science: serial, parallel, and distributed computing.

Choose Your Learning Path

🟡 Interactive Warehouse Challenge

Duration: 5–10 minutes | Level: Beginner-Friendly

Hands-on game where you manage delivery lanes to understand parallel processing vs. sequential work. Add and remove lanes (workers) to see how parallelism speeds up tasks, deliver packages within time limits, and physically see the difference between 1 lane vs. multiple lanes running simultaneously.

Best for: Visual learners who want immediate intuition

→ Start the Warehouse Challenge

🔵 Parallel & Distributed Computing Concepts

Duration: 15–20 minutes | Level: Intermediate

Deep dive into the theory and practice of serial, parallel, and distributed computing. Learn why 1 task at a time is slow, how multiple cores work together, and how many machines collaborate over a network. Includes real Python code examples you can modify and run. Also covers scenario challenges and AP CSP-style practice questions.

Best for: Students preparing for AP CSP exams or wanting hands-on Python code

→ Explore Parallel & Distributed Computing

🟢 Distributed Systems Deep Dive

Duration: 20–30 minutes | Level: Advanced

For those ready to go deeper: MapReduce pattern, how Google processes petabytes of data, fault tolerance (what happens when nodes crash!), and interactive visualizations of distributed systems in action. Includes real-world applications from Netflix, Google, and AI training platforms.

Best for: Advanced learners and those interested in distributed systems architecture

→ Explore Distributed Systems

Key Concepts at a Glance

Concept What It Does When to Use It Example
Serial Computing One task completes before the next starts When tasks are dependent or speedup doesn’t matter Processing a single image
Parallel Computing Multiple cores on ONE machine work simultaneously When tasks are independent and data fits in RAM Rendering 16 video frames on an 8-core laptop
Distributed Computing Many SEPARATE machines collaborate over a network When data is too large for one machine or scale is critical Training an AI model on 10,000 GPUs across data centers

Learning Outcomes

After completing this lesson hub, you’ll be able to:

  • ✅ Identify which computing paradigm is right for a given problem
  • ✅ Calculate speedup and understand Amdahl’s Law
  • ✅ Explain why distributed systems need fault tolerance
  • ✅ Recognize real-world applications of parallel and distributed computing
  • ✅ Write parallel Python code using ThreadPoolExecutor
  • ✅ Design solutions that scale from laptops to data centers

If you’re new to computing:

  1. Start with the Warehouse Challenge for intuition
  2. Move to Parallel & Distributed Computing for code examples
  3. Jump to Distributed Systems if you’re curious about scale

If you’re preparing for AP CSP:

  1. Focus on Parallel & Distributed Computing — this covers exam topics
  2. Review the scenario challenges and practice questions
  3. Reference Distributed Systems for bonus depth

If you’re a visual learner:

  1. Warehouse Challenge first
  2. Then the theory in Parallel & Distributed Computing

Common Misconceptions

"More cores = always faster"

❌ False. Tasks must be parallelizable. Some work (like updating a single counter) can't be sped up by adding cores. See Amdahl's Law.

"Distributed = always better than parallel"

❌ False. Distributed adds network overhead. Parallel is faster when all data fits on one machine. Use distributed only when you need the scale.

"If one node crashes, the whole system fails"

❌ False. Fault-tolerant distributed systems automatically reassign work to healthy nodes. The job keeps running.

Quick Reference: When to Use What

Is the data larger than 1 machine can hold?
  → YES: Use DISTRIBUTED computing
  → NO: Continue below

Do all tasks fit on 1 machine but take too long?
  → YES: Use PARALLEL computing (multiple cores)
  → NO: Use SERIAL computing (1 core is fine)

Are tasks fully independent?
  → YES: Parallelism will help
  → NO: Sequential parts will limit speedup

Resources

Ready to start? Pick a lesson above and jump in! 🚀