Statistically Analyzing Execution Variance for Soft Real-Time Applications

Kumar, T. K. Satish; Cledat, Romain; Sreeram, Jaswanth; Pande, Santosh

doi:10.1007/978-3-540-89740-8_9

Cited by 4 publications

(3 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In other words, the stage assignment, workload distribution and scheduling are planned before the program's execution commences. However, a lot of streaming applications show ample variations in execution time [11,16], especially with recent multimedia standards where input streams may contain a lot of varying values, types and rates of data being presented to the application. Figure 1(a) shows an example of a stream graph with three partitions.…”

Section: Background and Motivationmentioning

confidence: 99%

“…However, they model only the effects of dynamic variations rather than the causes, making it infeasible to predict variation in advance by considering program inputs, which are the root cause. In [11], program execution is analyzed with different input sets to identify which parts of the program have the highest statistical variance in execution time, which is mainly geared towards aiding programmers. We go a step further in actually analyzing the program in such a way that the execution time and its variations can actually be estimated for a new input given to the program.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Input-driven dynamic execution prediction of streaming applications

Aleen

Sharif

Pande

2010

Proceedings of the 15th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

Self Cite

View full text Add to dashboard Cite

Streaming applications are promising targets for effectively utilizing multicores because of their inherent amenability to pipelined parallelism. While existing methods of orchestrating streaming programs on multicores have mostly been static, real-world applications show ample variations in execution time that may cause the achieved speedup and throughput to be sub-optimal. One of the principle challenges for moving towards dynamic pipeline balancing has been the lack of approaches that can predict upcoming dynamic variations in execution efficiently, well before they occur.In this paper, we propose an automated dynamic execution behavior prediction approach that can be used to efficiently estimate the time that will be spent in different pipeline stages for upcoming inputs without requiring program execution. This enables dynamic balancing or scheduling of execution to achieve better speedup. Our approach first uses dynamic taint analysis to automatically generate an input-based execution characterization of the streaming program, which identifies the key control points where variation in execution might occur with the associated input elements that cause these variations. We then automatically generate a lightweight emulator from the program using this characterization that can predict the execution paths taken for new streaming inputs and provide an execution time estimates and possible dynamic variations. We present experimental evidence that our technique can accurately and efficiently estimate execution behaviors for several benchmarks. Our experiments show that dynamic pipeline balancing using our predicted execution behavior can achieve considerably higher speedup than static balancing.

show abstract

Section: Background and Motivationmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Input-driven dynamic execution prediction of streaming applications

Aleen

Sharif

Pande

2010

Proceedings of the 15th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

Self Cite

View full text Add to dashboard Cite

show abstract

“…A domain-agnostic approach is to statistically analyze the execution variance of soft-real time applications (Kumar et al, [11]). Using profiling, components of the applications that lead to variable execution times are first identified.…”

Section: Related Workmentioning

confidence: 99%

Stochastic modeling and performance analysis of multimedia SoCs

Raman

Nouri

Gangadharan

et al. 2013

2013 International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation (SAMOS)

View full text Add to dashboard Cite

Abstract-Quality of video and audio output is a design-time constraint for portable multimedia devices. Unfortunately, there is a huge cost (e.g. buffer size) incurred to deterministically guarantee good playout quality; the worst-case workload and the timing behavior can be significantly larger than the average-case due to high variability in a multimedia system. In future mobile devices, the playout buffer size is expected to increase, so, buffer dimensioning will remain as an important problem in system design. We propose a probabilistic analytical framework that enables low-cost system design and provides bounds for playing acceptable multimedia quality. We compare our approach with a framework comprising both simulation and statistical model checking, built to simulate large embedded systems in detail. Our results show significant reduction in output buffer size compared to deterministic frameworks.

show abstract

Characterizing Dominant Program Behavior Using the Execution-Time Variance of the Call Structure

Kumar

Pande

2019

2019 IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS)

View full text Add to dashboard Cite

Statistically Analyzing Execution Variance for Soft Real-Time Applications

Cited by 4 publications

References 12 publications

Input-driven dynamic execution prediction of streaming applications

Input-driven dynamic execution prediction of streaming applications

Stochastic modeling and performance analysis of multimedia SoCs

Characterizing Dominant Program Behavior Using the Execution-Time Variance of the Call Structure

Contact Info

Product

Resources

About