Explaining C++ memory ordering (frequent interview topic btw)

3-Second Blueprint (English)

1. Relaxed atomic operations guarantee single-variable atomicity without ordering prior writes, so reader threads may still observe stale data.
2. Pairing a release store with an acquire load synchronizes two threads, guaranteeing all writes before the release become visible afterward.
3. Sequential consistency enforces one global modification order across every thread at the hardware cost of extra cache-coherency traffic.