Sequential guarantees

Understanding sequential consistency in C++ memory ordering

This is an easy conceptual question on C++ concurrency that tests your grasp of memory ordering guarantees — a foundational topic for anyone writing lock-free code or working with concurrent data structures. It appears frequently in systems and trading-systems interviews where correctness under concurrency is non-negotiable.

The question focuses on one specific memory ordering mode and what invariants it enforces across multiple threads. To answer it well, you need to distinguish between different orderings (relaxed, acquire/release, sequential consistency) and understand what constraints each imposes on instruction reordering and visibility across threads. The answer should be precise about both compile-time and runtime guarantees.

• Happens-before relationships and synchronization points
• Load and store ordering constraints
• Trade-offs between correctness and performance