ETD

• energy-efficient algorithms
• parallel algorithms
• Rust
• suffix array construction
• suffix arrays

This thesis investigates the energy efficiency of suffix array construction algorithms (SACAs), analyzing how algorithmic design, programming language, and parallelism influence both performance and energy consumption. Several SACAs are experimentally evaluated using an extended version of the SacaBench framework integrated with hardware energy measurements through Intel Running Average Power Limit (RAPL). As a design contribution and to further investigate the impact of parallelism, a parallel extension of the Larsson-Sadakane qsufsort algorithm, namely PQSufsort, is introduced. The algorithm is implemented in both Rust and C, enabling an evaluation of how programming-language abstractions, memory-safety guarantees, and parallel scheduling strategies affect runtime performance, scalability, and energy consumption on large datasets.