An industrial design space exploration framework for supporting run-time resource management on multi-core systems

Abstract

Current multi-core design methodologies are facing increasing unpredictability in terms of quality due to the actual diversity of the workloads that characterize the deployment scenario. To this end, these systems expose a set of dynamic parameters which can be tuned at run-time to achieve a specified Quality of Service (QoS) in terms of performance. A run-time manager operating system module is in charge of matching the specified QoS with the available platform resources by manipulating the overall degree of task-level parallelism of each application as well as the frequency of operation of each of the system cores.In this paper, we introduce a design space exploration framework for enabling and supporting enhanced resource management through software re-configuration on an industrial multicore platform. From one side, the framework operates at design time to identify a set of promising operating points which represent the optimal trade-off in terms of the target power consumption and performance. The operating points are used after the system has been deployed to support an enhanced resource management policy. This is done by a light-weight resource management layer which filters and selects the optimal parallelism of each application and operating frequency of each core to achieve the QoS constraints imposed by the external world and/or the user.We show how the proposed design-time and run-time techniques can be used to optimally manage the resources of a multiple-stream MPEG4 encoding chip dedicated to automotive cognitive safety tasks 1 .

Effective Runtime Resource Management Using Linux Control Groups with the BarbequeRTRM Framework

“…A single OP is characterized by a set of application-specific parameters affecting either the quality of the computation or the application performance, but characterized by values ranging only in the boundaries defined by the current set of assigned resources (AWMs).The set of application configurations is defined via a design-time profiling that is supported by design space exploration (DSE) tools. The effectiveness of these techniques has been already proved by prior works (e.g., Yang and Catthoor [2003], Ykman-Couvreur et al [2005], and Mariani et al 2010]) that show how DSE is helpful in supporting runtime resource management.…”

A hierarchical run-time adaptive resource allocation framework for large-scale MPSoC systems

“…Assuming a sequence of application scenarios need to be execution on the target system, our objective is to minimize the total execution time of the system, in other words, to maximize the throughput for target application scenarios. are usually considered because of their effectiveness and simplicity [18]. However, when the system scales, a centralised approach often suffers from its performance bottleneck as heavy communication might happen during resource reallocation when the number of processing elements is very large.…”

“…Among these research, the hybrid task mapping approach is commonly used which combines the design-time preparation with the run-time dynamic mapping policy to do task reallocation. For example, Mariani et al [17] proposed a run-time management framework in which Pareto-fronts with system configuration points for different applications are determined during design-time DSE, after which heuristics are used to dynamically select a proper system configuration at run time. In [45], a fast and light-weight priority based heuristic is used to select near-optimal configurations explored at design time for the active applications according to the available platform resources.…”

“…This framework takes advantages of the state-of-the-art solutions for modern adaptive heterogeneous MPSoC systems where the system adaptivity is achieved by adaptively adjusting the mapping of applications to the underlying hardware resources which is optimised at design time. More specifically, we reconsider the method in the state-of-the-art solutions that allows for dynamically reconfiguring the system at run time based on pre-optimized system configurations, such as task mappings derived at design time [17,25,27,29,31,37,45], and extend it by solving the issues of scalability as introduced in [25] and blind adaptivity (a system reconfiguration that should not have happened because of its large overhead actually happened) that are usually existed in these solutions for our target large-scale heterogeneous MPSoC system.…”

Runtime adaptive allocation of dynamically mixed tasks on a heterogeneous MPSoC platform

“…One possible improvement of pure static approaches is to use semi-static scenario-based algorithms [8], [9], [10], [11], [12]. The basic idea is to compute a separate schedule for each possibly coexisting application combination and store all configuration data in a table.…”