Spin wars

What this spinlock concurrency question tests

This is an easy concurrency question that evaluates your understanding of busy-waiting and the performance characteristics of spinlocks under contention. It's common in interviews at firms where low-latency C++ code and thread synchronization are fundamental.

The question asks you to reason about how different spinlock implementations behave when many threads compete for the same lock simultaneously. Rather than memorizing API details, you'll need to think through CPU cache coherency, memory bandwidth, and the cost of failed spin attempts. Strong candidates can explain the trade-off between spinning (active waiting) and blocking, and predict which approach will waste fewer resources or cause more contention on the memory bus.

• Spin vs. block semantics and their performance trade-offs
• Cache coherency and false sharing in high-contention scenarios
• Memory ordering and volatile access patterns
• Thread wake-up latency and power efficiency