Why data layout matters more than asymptotic time complexity for CPU cache performance:

• Array of Structures (AoS): struct Particle { float x, y, z; int id; char name[32]; }; std::vector<Particle> particles;
  Iterating over positions loads 48 bytes per particle into CPU cache lines, wasting 75% bandwidth on unused name data!
• Structure of Arrays (SoA): struct ParticleSystem { std::vector<float> posX, posY, posZ; std::vector<int> id; };
  Positions are contiguous in RAM. CPU prefetcher loads 16 consecutive float coordinates per 64-byte cache line, enabling SIMD vectorization and 3.8x faster execution!

SoA layout is the foundational principle behind Data-Oriented Design (DOD) and Entity Component Systems (ECS).

Cache misses cost 100-300 CPU cycles each. Keeping hot data packed tightly in memory is how you write real-time software.