What coding choices impact performance?
Performance is affected by: • The way loops are organized. For best performance, the innermost loop in a loop nest should access arrays with a stride of one. • The way the data is structured. For example, a single structure containing arrays of data may be more efficient than an array of structures, since the access to the arrays may allow vectorization. • The iteration counts of loops. Longer iteration counts are generally better, as the loop overhead is amortized over more iterations. Tiny iteration counts (like two or three elements) may be faster to process in scalar. Extremely long loops (fluxing through tens of thousands of array elements) can exceed the size of cache and interfere with data reuse. • The data type of arrays. For example, if the arrays are double precision floating point, then you would get no speedup from AltiVec. • How the memory hierarchy is used. Most current processors are relatively unbalanced between memory bandwidth and CPU capacity. If you have large data
Related Questions
- What kind of impact will a Unitrends DPU have on my server performance while its running a Hot Bare Metal or Cold Bare Metal backup routine?
- How will Colombian economic policy-making and performance impact on corporate profitability over 2009-2014?
- What is the performance impact of TurboGold on a Domino server?
