Volume 5: Hands-On Labs

The exercises in Volume 5 come in three tiers, going from easier to harder and from individual building blocks to complete systems.

Tier one is made up of the small in-article exercises attached to the end of each article, which verify a single knowledge point — for example, recognizing a data race, predicate-based waiting on a condition_variable, or picking the right memory order for atomics. Each exercise takes 10–20 minutes and needs no extra project scaffolding.

Tier two is the chapter-scale assignments (Labs): the 8 labs listed on this page. Each lab is a runnable mini system broken into 3–5 milestones, and every milestone has a well-defined interface, Catch2 tests, and acceptance criteria. When you finish one, you should walk away with a reusable concurrency component, not a pile of scattered demos.

Tier three is the end-of-volume capstone project (Capstone), which wires the components from all the earlier labs together into a mini concurrent runtime.

Lab Overview

Lab	Project Name	Chapters Covered	Suggested Time	Difficulty	Prerequisite Labs
Lab 0	Thread Lifecycle	ch00–ch01	4–6h	intermediate	None
Lab 1	Bounded Queue & Sync Primitives	ch02–ch04	8–12h	intermediate	Lab 0
Lab 2	Atomic Metrics & SPSC Ring Buffer	ch03–ch04	6–8h	intermediate	Lab 0
Lab 2.5	Concurrency Debugging	ch08	3–4h	intermediate	Lab 0–2
Lab 3	Production-style Thread Pool	ch05	10–14h	advanced	Lab 0–1
Lab 4	Coroutine Scheduler & Event Loop	ch06	12–16h	advanced	Lab 3
Lab 5	Channel or Actor Runtime	ch07	8–12h	advanced	Lab 1, 4
Capstone	Mini Concurrent Runtime	ch08–ch09	8–12h	advanced	Lab 0–5

The bare minimum — completing Lab 0, Lab 1, Lab 3, and the Capstone (about 30–45 hours) — already covers the core competency curve of Volume 5. Working through every lab in full takes roughly 60–85 hours.

Lab Dependencies

Lab 0 (joining_thread / thread_guard)
  │
  ├─→ Lab 1 (BoundedBlockingQueue, ConcurrentCache)
  │     │
  │     ├─→ Lab 2 (SpscRingBuffer)        ─ independent implementation, does not depend on Lab 1
  │     │
  │     ├─→ Lab 2.5 (Debugging Lab)        ─ reuses code from Labs 0–2 as diagnostic material
  │     │
  │     └─→ Lab 3 (ThreadPool)             ─ reuses Lab 1's BoundedBlockingQueue
  │           │
  │           ├─→ Lab 4 (Coroutine Scheduler) ─ shutdown semantics follow Lab 3
  │           │     │
  │           │     └─→ Lab 5 (Channel/Actor) ─ can reuse Lab 1's queue
  │           │
  │           └─→ Capstone (Mini Runtime)   ─ composes the components of Labs 0–5

Environment Setup

All labs share the following environment requirements:

• Compiler: GCC 12+ or Clang 15+ (C++20, full coroutine support)
• CMake: 3.14+
• Test framework: Catch2 v3 (header-only, fetched via FetchContent)
• TSan: compile flags -fsanitize=thread -g
• Platform: Linux or WSL2 (required for the epoll part of Lab 4)
• Valgrind (optional; needed for helgrind in Lab 2.5)

Every lab article opens with a concrete CMakeLists.txt template that you can use directly.