Structures for phase classification

Lau, Jeremy; Schoemackers, S.; Calder, Brad

doi:10.1109/ispass.2004.1291356

Cited by 75 publications

(68 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…It examines the application's execution path to detect hardware independent phases [21,14]. Such phases can be readily missed by performance counter based phase detection, while changes in executed code reflect changes in many different metrics [20,21,5,22,9,18]. To leverage this, ScarPhase monitors what code is executed by dividing the application into windows and using hardware performance counters to sample which branches execute in a window.…”

Section: Phase Detectionmentioning

confidence: 99%

Modeling performance variation due to cache sharing

Sandberg

Sembrant

Hägersten

et al. 2013

2013 IEEE 19th International Symposium on High Performance Computer Architecture (HPCA)

View full text Add to dashboard Cite

Shared cache contention can cause significant variability in the performance of co-running applications from run to run. This variability arises from different overlappings of the applications' phases, which can be the result of offsets in application start times or other delays in the system. Understanding this variability is important for generating an accurate view of the expected impact of cache contention. However, variability effects are typically ignored due to the high overhead of modeling or simulating the many executions needed to expose them. This paper introduces a method for efficiently investigating the performance variability due to cache contention. Our method relies on input data captured from native execution of applications running in isolation and a fast, phaseaware, cache sharing performance model. This allows us to assess the performance interactions and bandwidth demands of co-running applications by quickly evaluating hundreds of overlappings.We evaluate our method on a contemporary multicore machine and show that performance and bandwidth demands can vary significantly across runs of the same set of co-running applications. We show that our method can predict application slowdown with an average relative error of 0.41% (maximum 1.8%) as well as bandwidth consumption. Using our method, we can estimate an application pair's performance variation 213× faster, on average, than native execution.

show abstract

Section: Phase Detectionmentioning

confidence: 99%

Modeling performance variation due to cache sharing

Sandberg

Sembrant

Hägersten

et al. 2013

2013 IEEE 19th International Symposium on High Performance Computer Architecture (HPCA)

View full text Add to dashboard Cite

show abstract

“…based on simulation results which often correspond to small performance differences. Recent research works on sampling have similar accuracy targets [6,7,13].…”

Section: Introduction and Related Workmentioning

confidence: 99%

“…While the recent surge of research articles on sampling started with rather large sample sizes (100M in the first SimPoint article [12]), it has later shifted to very small intervals (1,000 in SMARTS [14]), and it is now converging to intermediate sizes (1M and 10M in SimPoint [10,6]), and even to varying sizes in EXPERT [7] and SimPoint VLI [5] (ranging from 52,000 to 6.1M in EXPERT, and 100M to 500M in SimPoint VLI). With 100M samples, warm-up is not an issue, at least with current cache sizes.…”

Section: Introduction and Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Budgeted region sampling (BeeRS): do not separate sampling from warm-up, and then spend wisely your simulation budget

Pérez

Berry

Temam

2005

Proceedings of the Fifth IEEE International Symposium on Signal Processing and Information Technology, 2005.

View full text Add to dashboard Cite

Abstract. While the recent surge of research articles on sampling started with rather large sample sizes, it has later shifted to very small intervals, and it is now converging to intermediate sizes, and even to varying sizes. With 100M samples, warm-up is not an issue, at least with current cache sizes. However, with significantly smaller samples, warm-up becomes critical, especially when the sampling target accuracy is of the order of a few percent. However, in most sampling research works, warm-up has largely been treated as a separate issue.In this article, we advocate for an integrated approach at (simulator-based) warm-up and sampling. Instead of separating warm-up and sampling, we take exactly the opposite approach, provide a common instruction budget for warm-up and sampling, and we attempt to spend it as wisely as possible on either one.This budget and integrated approach at warm-up and sampling achieves an average CPI error of 1.68% on the 26 Spec benchmarks with an average sampling size of 288 millions instructions, and at the same time, it relieves the user from any delicate decision such as setting the sampling or warm-up sizes, thanks to the integrated warm-up+sampling and the region partitioning approaches.

show abstract

“…Two pervious works were in representative sampling. Lau et al [5] extensively examined different phase features based on data accesses. Their features are related to the access patterns, but do not directly measure the data locality.…”

Section: Reuse Distance Distribution (Rdd) Definitionmentioning

confidence: 99%