Lock-Free Queue Benchmarks

A C++ benchmark suite comparing lock-free and mutex-based queue implementations under multi-threaded workloads.
Includes SPSC (single-producer single-consumer) queues implementation, built from scratch to explore performance trade-offs in concurrent data structures.

---

Implementations

Queue Type	Description
SPSCQueue	Lock-free single-producer, single-consumer ring buffer. No CAS needed; minimal atomics.
MutexBoundedQueue	Same ring-buffer layout but protected by a single std::mutex for fairness comparison.

All queues share the same buffer structure and try_push / try_pop API.

---

Benchmark Overview

Each benchmark runs for ~10 seconds, measuring total throughput and item count.

• Producers: generate integers continuously until stop signal.
• Consumer: drains the queue as fast as possible.
• Result: total produced/consumed items and throughput (million items/sec).

Running Lock-Free SPSC Queue benchmark, capacity=1048576, time=10s

Running Mutex-SPSC Queue benchmark, capacity=1048576, time=10s

Example Output

=== Results ===
Lock-free SPSC:
  time:       10.2046 s
  produced:   83128472
  consumed:   83128472
  throughput: 8.14617 M items/s
Mutex SPSC:
  time:       10.202 s
  produced:   59115215
  consumed:   59115215
  throughput: 5.79447 M items/s