The increasing availability of different kinds of processing resources in heterogeneous system architectures associated with today’s fast-changing, unpredictable workloads has propelled an interest towards systems able to dynamically and autonomously adapt how computing resources are exploited to optimize a given goal.
Self-adaptiveness and hardware-assisted virtualization are the two key-enabling technologies for this kind of architectures, to allow the efficient exploitation of the available resources based on the current working context.
The EU FP7 project SAVE (Self-Adaptive Virtualization-aware high-performance/low-Energy heterogeneous system architectures ) aims at addressing the challenge of exploiting specialized computing resources of a heterogeneous system architecture (HSA) by pooling them and taking advantage of their individual characteristics to optimize the performance/energy trade-off for the resulting system, without constraining the applications or operation context.
To this end, we strive for defining a more general approach for exploiting HSAs, low- ering the complexity of managing the available resources, while enabling the over- all system to pursue an optimization goal that can depend on the current working conditions (application requirements, workload, …).
More precisely, SAVE will address these limitations by providing self-adaptivity and hardware-assisted vir- tualization to allow the system to dynamically and autonomously decide how to optimally allocate the workload generated by applications to the specialized resources for achieving an effective execution of the application while optimizing a user-defined goal (e.g., performance, energy, reliability, resource utilization).
SAVE will define the necessary SW/HW technologies for implementing self- adaptive systems exploiting heterogeneous architectures that include two classes of accelerators: GPUs and DFEs that enhance heterogeneous architectures to cope with the increased variety and dynamics of workloads observed in the HPC and ES domains.
Virtualization and self-adaptiveness are jointly exploited to obtain a new self-adaptive virtualization-aware Heterogeneous System Architec- ture, dubbed saveHSA, that exhibits a highly adaptive behavior to achieve the requested performance while minimizing energy consumption by allocating the tasks to the most appropriate accelerators, based on the current status of the overall system.
The effectiveness of SAVE’s technologies will be validated in two applications scenarios, financial risk computing and image processing algorithms, to cover the ES and HPC domain. We strive for an energy-efficiency improve- ment of 20% with respect to today’s architecture that use DFEs or GPUs in a traditional fashion. At the same time, system manageability, ease of deployment and resilience will be greatly improved.
(***Other authors of this paper are M. Coppola, STMicroelectronics; K. Djafarian, ARM; G. Kornaros. Technological Educational Institute of Crete; M. Paolino, Virtual Open Systems; O. Pell, Maxeler Technologies; and C. Plessl, University of Paderborn, Germany. )
To read this external content in full, download the complete paper in PDF form from the open online author archives at the SAVE Project.