Static Probabilistic Timing Analysis of Random Replacement Caches using Lossy Compression

Griffin, David; Lesage, Benjamin; Burns, Alan; Davis, Robert I.

doi:10.1145/2659787.2659809

Cited by 15 publications

(11 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach is improved in , with an improved algorithm for SPTA. [Griffin et al 2014] propose a methodology from the field of Lossy Compression and they use a fully-associative cache for timing analysis throughout their work. By using May and Must Analysis, the result is more accurate with appropriate parameters.…”

Section: Related Workmentioning

confidence: 99%

An Adaptive Markov Model for the Timing Analysis of Probabilistic Caches

Chen

Beltrame

2017

ACM Trans. Des. Autom. Electron. Syst.

View full text Add to dashboard Cite

Accurate timing prediction for real-time embedded software execution is becoming a problem due to the increasing complexity of computer architecture, and the presence of mixed-criticality workloads. Probabilistic caches were proposed to set bounds to Worst Case Execution Time (WCET) estimates and help designers improve real-time embedded system resource use. Static Probabilistic Timing Analysis (SPTA) for probabilistic caches is nevertheless difficult to perform, because cache accesses depend on execution history, and the computational complexity of SPTA makes it intractable for calculation as the number of accesses increases. In this paper, we explore and improve SPTA for caches with evict-on-miss random replacement policy using a state space modeling technique. A nonhomogeneous Markov model is employed for single-path programs in discrete-time finite state space representation. To make this Markov model tractable, we limit the number of states and use an adaptive method for state modification. Experiments show that compared to the state-of-the-art methodology, the proposed adaptive Markov chain approach provides better results at the occurrence probability of 10 −15 : in terms of accuracy, the state-of-the-art SPTA results are more conservative, by 11% more on average. In terms of computation time, our approach is not significantly different from the state-of-the-art SPTA.

show abstract

Section: Related Workmentioning

confidence: 99%

An Adaptive Markov Model for the Timing Analysis of Probabilistic Caches

Chen

Beltrame

2017

ACM Trans. Des. Autom. Electron. Syst.

View full text Add to dashboard Cite

show abstract

“…For example, by replacing the traditional LRU or pseudo-LRU cache replacement policy for a policy that randomly chooses the cache line to be evicted (and assume that every cache line has the same probability to get evicted), the time overhead due to cache penalties and cache line evictions can be analysed as an i.i.d. random variable with a known distribution (see [22], [23], [24] for examples of such techniques). If every source of interference exhibits a randomized behavior with a known distribution then the execution time itself can be analysed statically.…”

Section: Probabilistic Techniquesmentioning

confidence: 99%

Methodologies for the WCET Analysis of Parallel Applications on Many-Core Architectures

Nélis

Yomsi

Pinho

2015

2015 Euromicro Conference on Digital System Design

View full text Add to dashboard Cite

There is an increasing eagerness to deploy and execute parallel applications on many-core infrastructures, preserving the time-predictability of the execution as required by real-time practices to upper-bound the response time of the embedded application. In this context, the paper discusses the application of the currently-available WCET analysis techniques and tools on such platforms and with highly parallel activities. After discussing the pros and cons of all different methodologies for WCET analysis, we introduce a new approach that is developed within the P-SOCRATES project. Abstract-There is an increasing eagerness to deploy and execute parallel applications on many-core infrastructures, preserving the time-predictability of the execution as required by real-time practices to upper-bound the response time of the embedded application. In this context, the paper discusses the application of the currently-available WCET analysis techniques and tools on such platforms and with highly parallel activities. After discussing the pros and cons of all different methodologies for WCET analysis, we introduce a new approach that is developed within the P-SOCRATES project.

show abstract

“…Lossy Compression, previously used on models of real time systems by Griffin et al, has been applied to static analyses of PLRU caches [16,18] and random replacement caches [17]. Given that Lossy Compression is capable of simplifying the models used in static analysis without discarding useful information, the same principles can be used to simplify a model populated by data from measurements.…”

Section: Related Workmentioning

confidence: 99%

“…To compensate for a lack of data, Lossy Compression [18] is applied to the model. Lossy compression states that in order to compress a model, one should first list the types of information contained in the model, evaluate their importance to the desired result, and discard or approximate information which is of low importance to the overall result.…”

Section: Compressing the Markov Chain Modelmentioning

confidence: 99%

“…Further, while previous applications of Lossy Compression were to Static Analysis [18], there is an additional benefit to applying Lossy Compression to a Measurement Based technique: as the Practically speaking, compressing the model such that the evidence supports the model is accomplished by modifying the question that is being asked of the model. Hence, instead of attempting to find an exact model of the system, for which there is insufficient evidence and therefore there is little confidence in the correctness of the model, an approximation is found.…”

Section: Compressing the Markov Chain Modelmentioning

confidence: 99%

See 1 more Smart Citation

Modelling fault dependencies when execution time budgets are exceeded

Griffin

Lesage

Bate

et al. 2015

Proceedings of the 23rd International Conference on Real Time and Networks Systems

Self Cite

View full text Add to dashboard Cite

Given that real-time systems are specified to a degree of confidence, budget overruns should be expected to occur in a system at some point. When a budget overrun occurs, it is necessary to understand how long such a state persists, in order to determine if the fault tolerance of the system is adequate to handle the problem. However, given the rarity of budget overruns in testing, it cannot be assumed that sufficient data will be available to build an accurate model. Hence this paper presents a new application of Markov Chain based modelling techniques combined with forecasting techniques to determine an appropriate fault model, using Lossy Compression to fit the model to the available data. In addition, a new algorithm, DepET, for generating job execution times with dependencies is given for use in task simulators.

show abstract

Static Probabilistic Timing Analysis of Random Replacement Caches using Lossy Compression

Cited by 15 publications

References 8 publications

An Adaptive Markov Model for the Timing Analysis of Probabilistic Caches

An Adaptive Markov Model for the Timing Analysis of Probabilistic Caches

Methodologies for the WCET Analysis of Parallel Applications on Many-Core Architectures

Modelling fault dependencies when execution time budgets are exceeded

Contact Info

Product

Resources

About