Games Engineering Optimisations

Games Engineering Optimisations

• The presentation covers various optimisation techniques, categorized as algorithmic, code-level, system-level, and parallel computing.
• Timing code is used to demonstrate optimisation effectiveness, including timing loops with conditional statements.
• Algorithmic optimisation involves choosing or developing the correct algorithm, understanding its complexity, and ensuring its suitability for the task. Suitable data structures are also crucial, taking into account data locality. Memoisation is a technique for optimizing algorithms prone to redundant calculations. A Fibonacci sequence calculation provides an example of this, demonstrating both recursive and memoised approaches. Notably, recursive solutions may have significant overhead due to repetitive calculations.

Optimisations (Further Detail)

• Algorithmic:

  • Choosing the appropriate algorithm is fundamental.
  • Understanding algorithm complexity is crucial.
  • Ensuring the algorithm matches the task requirements.
  • Using suitable data structures is important to optimise performance, notably considering locality properties.
  • Memoisation, a technique for optimizing recursive algorithms by storing the results of subproblems.

• Code-level:

  • Using compiler optimisations, and exploring other options available.
  • Using inline functions, enabling calculations at compile time.
  • Loop unrolling for improving loop performance.
  • Optimizing use of conditionals for reduced computational overhead.

• System-level:

  • Pipelining: The process of dividing tasks into multiple stages to execute concurrently; however, branching can impact efficiency.
  • Caching: Data and instruction caching improves speed by storing frequently accessed data closer to the processor. Policies such as direct mapping and set-associative mapping handle memory in the cache, taking into account the amount of data available and possible replacement strategies. Cache coherence is vital for concurrent processes.
  • OS Interference: Operating system activities can impact performance; memory handling is a key factor.

• Parallel Computing: Not explicitly detailed in the provided text.

Examples

• Fibonacci sequence: A demonstration of recursive and memoised approaches to algorithm optimisation.
• Rod Cutting Problem: Illustrates dynamic programming, which can optimise calculations of a sequence of cuts to maximise value.
• Vector operations: Examples in code demonstrate applying optimisation techniques. Examples showcase different array traversals in C++, showing how memory access patterns can significantly impact performance (e.g., row-major vs. column-major).

Explore various optimisation techniques in games engineering through algorithmic, code-level, system-level, and parallel computing approaches. Learn about the importance of choosing suitable algorithms, understanding their complexity, and optimizing code with techniques like memoisation. This quiz provides insights into practical applications such as timing code and evaluating performance improvements.