Energy-Efficient Thermal-Aware Scheduling for RT Tasks Using TCPN

Rubio-Anguiano, L. ; Desirena-López, G. ; Ramírez-Treviño, A. ; Briz, J.L. (Universidad de Zaragoza)

Resumen: This work leverages TCPNs to design an energy-efficient, thermal-aware real-time scheduler for a multiprocessor system that normally runs in a low state energy at maximum system utilization but its capable of increasing the clock frequency to serve aperiodic tasks, optimizing energy, and honoring temporal and thermal constraints. An off-line stage computes the minimum frequency required to run the periodic tasks at maximum CPU utilization, the proportion of each task''s job to be run on each CPU, the maximum clock frequency that keeps temperature under a limit, and the available cycles (slack) with respect to the system with minimum frequency. Then, a Zero-Laxity online scheduler dispatches the periodic tasks according to the offline calculation. Upon the arrival of aperiodic tasks, it increases clock frequency in such a way that all periodic and aperiodic tasks are properly executed. Thermal and temporal requirements are always guaranteed, and energy consumption is minimized.

Idioma: Inglés

DOI: 10.1016/j.ifacol.2018.06.307

Año: 2018

Publicado en: IFAC PAPERSONLINE 51, 7 (2018), 236-242
ISSN: 2405-8963

Factor impacto SCIMAGO: 0.298 - Control and Systems Engineering (Q3)

Financiación: info:eu-repo/grantAgreement/ES/DGA/T48
Financiación: info:eu-repo/grantAgreement/EC/H2020/687698/EU/High Performance and Embedded Architecture and Compilation/HiPEAC
Financiación: info:eu-repo/grantAgreement/ES/MINECO/TIN2016-76635-C2-1-R

Tipo y forma: Artículo (PostPrint)

Área (Departamento): Área Arquit.Tecnología Comput. (Dpto. Informát.Ingenie.Sistms.)