Self-Adaptive QoS Management of Computation and Communication Resources in Many-Core SoCs

Ruaro, Marcelo; Jantsch, Axel; Moraes, Fernando

doi:10.1145/3328755

Cited by 4 publications

(13 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1:7 • Task profile: by combining the scheduling and task communication information, the LLM traces the task profile, classifying it as computation-intensive, communication-intensive, or hybrid [17]. • Application heartbeat: typical applications running in MCSoCs are cyclic, such as video processing.…”

Section: Low-level Monitor-llmmentioning

confidence: 99%

“…Therefore, there are also messages involved in such protocols that physically change the system. Such messages are out of the scope of this work being part of the adaptive protocols: DVFS [10], dynamic communication switching [4,16], and task migration and mapping [17].…”

Section: Exchanged Messagesmentioning

confidence: 99%

“…User task observes the profile and performance of real-time tasks, by receiving deadlines/latency misses, heartbeats, and profiling [7,17]. The User task is concerned with the runtime QoS fulfillment of real-time tasks (computation and communication).…”

Section: Observation Tasksmentioning

confidence: 99%

“…Computation-intensive tasks can be early migrated to free cores. Communication-intensive tasks can early receive a CS connection [17].…”

Section: Observation Tasksmentioning

confidence: 99%

See 3 more Smart Citations

Modular and Distributed Management of Many-Core SoCs

Ruaro

Sant'Ana

Jantsch

et al. 2020

ACM Trans. Comput. Syst.

View full text Add to dashboard Cite

Many-Core Systems-on-Chip increasingly require Dynamic Multi-objective Management (DMOM) of resources. DMOM uses different management components for objectives and resources to implement comprehensive and self-adaptive system resource management. DMOMs are challenging because they require a scalable and well-organized framework to make each component modular, allowing it to be instantiated or redesigned with a limited impact on other components. This work evaluates two state-of-the-art distributed management paradigms and, motivated by their drawbacks, proposes a new one called Management Application (MA) , along with a DMOM framework based on MA. MA is a distributed application, specific for management, where each task implements a management role. This paradigm favors scalability and modularity because the management design assumes different and parallel modules, decoupled from the OS. An experiment with a task mapping case study shows that MA reduces the overhead of management resources (-61.5%), latency (-66%), and communication volume (-96%) compared to state-of-the-art per-application management. Compared to cluster-based management (CBM) implemented directly as part of the OS, MA is similar in resources and communication volume, increasing only the mapping latency (+16%). Results targeting a complete DMOM control loop addressing up to three different objectives show the scalability regarding system size and adaptation frequency compared to CBM, presenting an overall management latency reduction of 17.2% and an overall monitoring messages’ latency reduction of 90.2%.

show abstract

Section: Low-level Monitor-llmmentioning

confidence: 99%

Section: Exchanged Messagesmentioning

confidence: 99%

Section: Observation Tasksmentioning

confidence: 99%

“…Computation-intensive tasks can be early migrated to free cores. Communication-intensive tasks can early receive a CS connection [17].…”

Section: Observation Tasksmentioning

confidence: 99%

See 2 more Smart Citations

Modular and Distributed Management of Many-Core SoCs

Ruaro

Sant'Ana

Jantsch

et al. 2020

ACM Trans. Comput. Syst.

View full text Add to dashboard Cite

show abstract

“…It is possible to classify SDN proposals in three categories (first column of Table I): (i) adoption of a centralized SDN (C-SDN) control [4], [6], [8]-[ll]; (ii) distributed NoC control implemented inside each router or operating system (OS) [12], [13]; (Hi) distributed SDN control (D-SDN), similar to C-SDN but with each Controller in charge of a cluster of resources instead to the whole system [7], Most works apply SDN without being concerned with a specific objective (generic in the third column of Table I). Other works adopt a single objective, like QoS [4], [11], security [6], and power [12]. A multi-objective SDN proposal is a gap observed in the literature, tackled in this work.…”

mentioning

confidence: 98%

Multiple-objective Management based on a Distributed SDN Architecture for Many-cores

Ruaro

Moraes

2020

2020 33rd Symposium on Integrated Circuits and Systems Design (SBCCI)

Self Cite

View full text Add to dashboard Cite

The management of many-core systems is evolving to meet multiple objectives simultaneously. The Software-Defined Networking (SDN) has benefits explored in recent works that point it as a candidate to address this requirement at the communication level, at the same time that promotes management flexibility and reduced hardware complexity. Most of the research in SDN for many-cores assumes a centralized SDN (C-SDN) Controller and single-objective management. This work proposes multi-objective management based on a distributed SDN (D-SDN) architecture (SELF-SDN). The management is selfadaptive, addressing QoS and fault-tolerance simultaneously at the communication level. Experiments targeting QoS show that SELF-SDN provides a reduced amount of latency misses (-67%) and fast reaction time (-49.6%) to recover the QoS constraints compared to a C-SDN approach. Fault-tolerance experiments highlight the simplicity of the SDN paradigm to recover from faults in the NoC, not requiring additional hardware. Results related to multi-objective management demonstrate the fast reaction time of SELF-SDN to recover the communication latency faced to QoS loss and faults, reducing, on average, in 43% the reaction time compared to a C-SDN approach.

show abstract

SLAQA

Roy

Devaraj²,

Sarkar

et al. 2021

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

Continuous demands for higher performance and reliability within stringent resource budgets is driving a shift from homogeneous to heterogeneous processing platforms for the implementation of today’s cyber-physical systems (CPSs). These CPSs are typically represented as Directed-acyclic Task Graph (DTG) due to the complex interactions between their functional components that are often distributed in nature. In this article, we consider the problem of scheduling a real-time application modelled as a single DTG, where tasks may have multiple implementations designated as quality-levels, with higher quality-levels producing more accurate results and contributing to higher rewards/Quality-of-Service for the system. First, we introduce an optimal solution using Integer Linear Programming (ILP) for a DTG with multiple quality-levels, to be executed on a heterogeneous distributed platform . However, this ILP-based optimal solution exhibits high computational complexity and does not scale for moderately large problem sizes. Hence, we propose two low-overhead heuristic algorithms called Global Slack Aware Quality-level Allocator ( G-SLAQA ) and Total Slack Aware Quality-level Allocator ( T-SLAQA ), which are able to produce satisfactorily efficient as well as fast solutions within a reasonable time. G-SLAQA , the baseline heuristic, is greedier and faster than its counter-part T-SLAQA , whose performance is at least as efficient as G-SLAQA . The efficiency of all the proposed schemes have been extensively evaluated through simulation-based experiments using benchmark and randomly generated DTGs. Through the case study of a real-world automotive traction controller , we generate schedules using our proposed schemes to demonstrate their practical applicability.

show abstract

Self-Adaptive QoS Management of Computation and Communication Resources in Many-Core SoCs

Cited by 4 publications

References 25 publications

Modular and Distributed Management of Many-Core SoCs

Modular and Distributed Management of Many-Core SoCs

Multiple-objective Management based on a Distributed SDN Architecture for Many-cores

SLAQA

Contact Info

Product

Resources

About