We discuss the immediate need, challenges, and opportunities to transition into greener cloud computing platforms. Actions must be taken not only by the academy and the industry, but also by computer science practitioners from several areas, such as information systems, scheduling theory, distributed systems, HPC, computer architecture, and approximate computing, to cite a few.
This study presents a comprehensive comparative analysis of compiler efficiency, with a focus on energy consumption metrics across diverse domains of high-performance computing. By rigorously evaluating the performance of GCC, Clang, and ICC, the research aims to elucidate which compilers excel in specific areas, thereby providing valuable insights for the strategic selection of these tools based on the unique requirements of various computational tasks. The findings reveal that, among the total energy consumed during the computations, GCC accounted for 33.23%, Clang for 36.01%, and ICC for 30.76%. Notably, ICC demonstrated superior energy efficiency, being 7.43% more efficient than GCC, while Clang was 8.35% less efficient. These results underscore the critical importance of selecting the appropriate compiler to optimize energy efficiency in high-performance computing environments.
Investigamos o consumo de energia em benchmarks OpenMP com o objetivo de otimizar o desempenho em hardware de alto desempenho. Utilizamos técnicas de regressão linear e polinomial para prever o consumo de energia. Os resultados mais promissores incluem um erro de previsão de apenas 0.85% para o benchmark LavaMD no conjunto de benchmarks Rodinia e 0.94% para o benchmark Cholesky no conjunto Polybench.