preview

What Is Adaptive Recon Gurable Multi-Core Architecture

Better Essays
Open Document

Adaptive Recon gurable Multi-core Schedul- ing System [ARMSS] We have been designed an ecient oine and online scheduler of adaptive multi-core architectures. We present an optimal scheduler for an ideal adaptive multi-core architecture that is not restricted by any physical or technological constraint. Next, we impose additional constraints to perform scheduling on a realistic adaptive multi-core, divided into oine and online schedulers. That algorithm uses tasks having high and low utilization that run in the same cluster. When task has low utilization nished excuitation then task has height utilization can increase it's cores by one within the same cluster. The maximum number of cores to run one task equal the number of that …show more content…

That core actually run another task has low utilization Ul. Uh and Ul that are collected together must be belong to di erent type (TLP , ILP), line 11. See [Figure 4c] and [Table 1] where T6[Uh] was chosen to run part of it on the core that run task T5 [UI]. According to constraint number c3, the algorithm should collects tasks [Uh, Ul] running in the same cluster, [Figure 4d] . If task Uh don't blow to any cluster then collecting Uh+Ul to new cluster that satisfy equation[2], line 12 to 30 , see [ Figure 4d] and [Table 1 where T7[Uh] collect with T4[Ul]. According to constraint number c4 maximum number of cores can assign to task equal four cores because our model maximum cores per cluster equal four. After increasing number of cores assigned to any task, it should recalculate new utilization using equation [3] to this task and update the utilization Array[]. After determine task pair with utilization [Uh, Ul] that insure all the above constrains, the algorithm will increase Uh's task running core by one core, see [ Figure 4d] the algorithm increase number of cores running task T7 to two cores, and task T6 to three cores. The nal result after applying ARMSS algorithm on tasks [T1 .. T8] [Table 1], is shown in [Table 2] where T6 , T5 , T3 and T8 were running on the same cluster#1. T7 and T4 were running in the same cluster#2. Table 1: Example of real Tasks

Get Access
Related
Get Access