We give examples of how these optimizations are applied in the two upcoming chapters, where we optimize the computation of random minimizers (Groot Koerkamp and Martayan 2025) and build a fast minimal perfect hash function data structure (Groot Koerkamp 2025b). Modern CPUs have execution pipelines that are hundreds of instructions long Thus, if one instruction is waiting for some data (from memory), the CPU will already start execution upcoming instructions. Additionally, one should avoid sequential memory accesses, where the result of memory read determines the location of a second access to memory, since these can not be executed in parallel.