This paper presents Big-LITTLE Constant Bandwidth Server (BL-CBS), adynamic partitioning approach to schedule real-time task sets in an energy-efficient way on multi-core platforms based on the ARM big.LITTLE archi-tecture. BL-CBS is designed as an on-line and adaptive scheduler, based ona push/pull architecture that is suitable to be incorporated in the currentSCHED_DEADLINE code base in the Linux kernel. It employs a greedyheuristic to dynamically partition the real-time tasks among the big andLITTLE cores aiming to minimize the energy consumption and the migra-tions imposed on the running tasks. The new approach is validated throughthe open-source RT-Sim simulator, which has been extended integrating anenergy model of the ODROID-XU3 board, fitting tightly the power consump-tion profiles for the big and LITTLE cores of the board. An extensive setof simulations have been run with randomly generated real-time task sets,leading to promising results.
Dynamic Partitioned Scheduling of Real-Time Tasks on ARM big.LITTLE Architectures
Agostino Mascitti
Software
;Tommaso CucinottaSupervision
;Mauro MarinoniSoftware
;Luca AbeniSupervision
2021-01-01
Abstract
This paper presents Big-LITTLE Constant Bandwidth Server (BL-CBS), adynamic partitioning approach to schedule real-time task sets in an energy-efficient way on multi-core platforms based on the ARM big.LITTLE archi-tecture. BL-CBS is designed as an on-line and adaptive scheduler, based ona push/pull architecture that is suitable to be incorporated in the currentSCHED_DEADLINE code base in the Linux kernel. It employs a greedyheuristic to dynamically partition the real-time tasks among the big andLITTLE cores aiming to minimize the energy consumption and the migra-tions imposed on the running tasks. The new approach is validated throughthe open-source RT-Sim simulator, which has been extended integrating anenergy model of the ODROID-XU3 board, fitting tightly the power consump-tion profiles for the big and LITTLE cores of the board. An extensive setof simulations have been run with randomly generated real-time task sets,leading to promising results.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.