Scheduling analysis integration for heterogeneous multiprocessor SoC

Richter, Kai; Racu, Razvan; Ernst, Rolf

doi:10.1109/real.2003.1253270

Cited by 39 publications

(48 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…5. The upper arrival curve can be expressed as follows [Richter et al, 2003b], Although the periodic burst arrival curve cannot be approximated as the minimum of a set of staircase functions, it can be equivalent to a special logic composition of staircase functions. As shown in Fig.…”

Section: Periodic Burst Eventsmentioning

confidence: 99%

Evaluation and Improvements of Runtime Monitoring Methods for Real-Time Event Streams

Huang

Chen

et al. 2016

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

Runtime monitoring is of great importance as a safeguard to guarantee the correctness of system runtime behaviors. Two state-of-the-art methods, dynamic counters and l-repetitive function, were recently developed to tackle the runtime monitoring for real-time systems. While both are reported to be efficient in monitoring the arbitrary events, the monitoring performance between them has not yet been evaluated. This article evaluates both methods in depth, to identify their strengths and weaknesses. New methods are proposed to efficiently monitor the many-to-one connections that are abstracted as AND and OR components on multiple inputs. Representative scenarios are used as our case studies to quantitatively demonstrate the evaluations. Both methods are implemented in hardware FPGA. The timing overhead and resource usages of implementing the two methods are evaluated.

show abstract

Section: Periodic Burst Eventsmentioning

confidence: 99%

Evaluation and Improvements of Runtime Monitoring Methods for Real-Time Event Streams

Huang

Chen

et al. 2016

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

show abstract

“…Other approaches [1], [11] are based on the propagation of sequences of events [12] and analyze the system through the real-time calculus [13]. To best of our knowledge, however, the correlation between activation of consecutive activation of the same pipeline is lost introducing some pessimism in the analysis.…”

Section: Related Workmentioning

confidence: 99%

“…As the application is distributed over several processing elements, components are of distributed nature as well. For example, this is the typical scenario in the automotive context [1], [2].…”

Section: Introductionmentioning

confidence: 99%

The Demand Bound Function Interface of Distributed Sporadic Pipelines of Tasks Scheduled by EDF

Serreli

Lipari

Bini

2010

2010 22nd Euromicro Conference on Real-Time Systems

View full text Add to dashboard Cite

Abstract-In distributed real-time embedded systems (DRE), it is common to model an application as a set of task chains. Each chain is activated cyclically and must complete before an end-to-end deadline. Each task of the chain is bound to execute on a particular processing element.The complexity of designing and analyzing a DRE can be reduced by applying a component-based methodology: each pipeline can be seen as a component with its temporal characteristic summarized in its interface. Analysis can be carried out in two different steps: 1) derivation of the temporal interface of a component pipeline; 2) analysis of the whole system by integrating the temporal interfaces of the components.In this paper, we propose to describe the temporal interface of a task pipeline by a set of demand bound functions (dbf), one per each node on which the pipeline executes, and we describe an algorithm for computing the dbfs. First, we show that the scenario of strictly periodic activations is not the worst when the pipelines are sporadically activated. Then, we propose an exact algorithm for computing the dbfs. We show by experimental analysis that the computation time of the algorithm on pipelines with reasonable size is below one second on common PCs. Finally, we estimate the pessimism introduced by our analysis with respect to holistic analysis by an extensive set of simulations. I. INTRODUCTIONToday's applications are often developed by different vendors, each one providing separate components. As the application is distributed over several processing elements, components are of distributed nature as well. For example, this is the typical scenario in the automotive context [1], [2].In the analysis of such a system it is of key importance to preserve the following properties: 1) each vendor provides only a synthetic information on the developed component (called component interface); 2) the integration of the components is made only on the information contained in the interface. In real-time systems, a component is equipped also with a temporal interface that contains information related to the amount of computational resource required by the component over time. The analysis is then performed in two steps: in the first step, each component is analyzed in isolation, summarizing its temporal behavior with a (possibly small) set of temporal parameters. Such temporal parameters will be part of the component interface along with the functional and behavioral parameters. In the second step (integration), we must verify that the overall system is schedulable by integrating the temporal interfaces derived in the previous step.

show abstract

“…Our work builds upon the system-level timing analysis methods for event streams studied in [3,4,31,32]. Although, these papers also deal with end-to-end delays experienced by event streams and the computation of maximum buffer fill levels, their focus is on multiprocessor architectures and the modeling of resource sharing.…”

Section: Previous Work On Timing Analysis Of Event Streamsmentioning

confidence: 99%

Cache-Aware Timing Analysis of Streaming Applications

Chakraborty

Mitra

Roychoudhury

et al. 2007

19th Euromicro Conference on Real-Time Systems (ECRTS'07)

View full text Add to dashboard Cite

Of late, there has been a considerable interest in models, algorithms and methodologies specifically targeted towards designing hardware and software for streaming applications. Such applications process potentially infinite streams of audio/video data or network packets and are found in a wide range of devices, starting from mobile phones to set-top boxes. Given a streaming application and an architecture, the timing analysis problem is to determine the timing properties of the processed data stream, given the timing properties of the input stream. This problem arises while determining many common performance metrics related to streaming applications and the mapping of such applications onto hardware architectures. Such metrics include the maximum delay experienced by any data item of the stream and the maximum backlog or the buffer requirement to store the incoming stream. Most of the previous work related to estimating or optimizing these metrics take a high-level view of the architecture and neglect micro-architectural features such as caches. In this paper, we show that an accurate estimation of these metrics, however, heavily relies on an appropriate modeling of the processor micro-architecture. Towards this, we present a novel framework for cacheaware timing analysis of stream processing applications. Our framework accurately models the evolution of the instruction cache of the underlying processor as a stream is processed, and the fact that the execution time involved in processing any data item depends on all the previous data items occurring in the stream. The main contribution of our method lies in its ability to seamlessly integrate program analysis techniques for micro-architectural modeling with known analytical methods for analyzing streaming applications, which treat the arrival/service of event streams as mathematical functions. This combination is powerful as it allows to model the code/cache-behavior of the streaming application, as well as the manner in which it is triggered by event arrivals. We employ our analysis method to an MPEG-2 encoder application and our experiments indicate that detailed modeling of the cache behavior is efficient, scalable and leads to more accurate timing/buffer size estimates.

show abstract

Scheduling analysis integration for heterogeneous multiprocessor SoC

Cited by 39 publications

References 27 publications

Evaluation and Improvements of Runtime Monitoring Methods for Real-Time Event Streams

Evaluation and Improvements of Runtime Monitoring Methods for Real-Time Event Streams

The Demand Bound Function Interface of Distributed Sporadic Pipelines of Tasks Scheduled by EDF

Cache-Aware Timing Analysis of Streaming Applications

Contact Info

Product

Resources

About