DataOriented

Data-oriented design is a design and programming paradigm that prioritizes the structure and flow of data over the organization of code into objects and methods. It emphasizes how data is laid out in memory and how it is accessed and processed by the CPU, with the goal of maximizing data locality, cache efficiency, and vectorization on modern architectures.

Key concepts include data layout strategies such as structure of arrays (SoA) versus array of structures (AoS), contiguous memory layouts, data-driven processing pipelines, and minimizing indirect memory access. In practice, data-oriented design separates data representation from the algorithms that operate on it and favors iterations over large, cache-friendly blocks of data.

It is widely applied in performance-critical domains, notably game development, high-performance computing, graphics, and simulation. It often leads to reorganizing systems around data processing rather than object hierarchies, and is commonly associated with patterns like entity-component-system architectures, where component data is stored contiguously and processed in tight loops.

Compared with object-oriented design, data-oriented design focuses on how data moves through the system and how it is laid out in memory, rather than on encapsulation and polymorphic behavior. This does not forbid abstraction, but encourages design decisions that reveal and exploit data access patterns.

Critics note that data-oriented design can increase code complexity and reduce readability if applied without careful profiling and clear goals; it is most effective when profiling shows data access as the bottleneck.