Context-aware cloud computing for HEP

Abstract

The use of cloud computing is increasing in the field of high-energy physics (HEP). Both dedicated and opportunistic cloud resources are delivering a growing fraction of the requirements of large projects. The use of context-aware designs, such as those used in the mobile communication industry, can improve our distributed systems and make them more responsive to changing demands and dynamic resources. We show how context-aware service have been implemented and their impact on the system. We describe our plans to add new context-aware capabilities in the future.

A holistic study of the WLCG energy needs for the LHC scientific program

“…Software releases currently exist for many of the experiment workflows, and most have now been validated for use in production. GridPP compared [8] the performance and the power consumption of two machines, roughly similar in terms of specifications and prices: The HEP community implemented the HEPScore [9] benchmark suite to replace HS06. The advantage of HEPScore is the use of the High Energy Physics experiments workflows as applications in the benchmark, giving a more realistic estimate of the performance of the processors for the HEP needs.…”

“…The workflows available at the time of this study in HEPscore were compiled and executed in the two architectures above. Detailed results are available in [8] but in summary: using hyperthreading on the x86 AMD processor gives gains both in terms of speed and energy efficiency. However, the gains are highly workload dependent.…”