9. SIMD Programming on the PPU, Part 2: Methods and Algorithms
The previous chapter discussed the AltiVec, SIMD Math, and MASSV libraries in great detail, but there’s more to SIMD programming than just knowing the functions. This chapter delves deeper into vector coding and describes methods that will improve the speed and efficiency of your applications. In particular, it covers two points:
- Accessing unaligned memory and unrolling loops
- Structure of arrays (SOA) versus array of structures (AOS)
This chapter also provides many real-world examples of vector computation on the PPU. This will not only give you a better understanding of AltiVec coding, but you’ll also have a set of common, high-performance vector utilities that you can use in your ...