Understanding transactional memory performance

Porter, Donald E.; Witchel, Emmett

doi:10.1109/ispass.2010.5452061

Cited by 13 publications

(14 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Usui et al [41] use a simple cost-benefit analysis to choose between locking and transactions. The performance model from [35] focuses on modeling transactional conflict behavior. Unlike estima, this approach requires heavy instrumentation of the application memory accesses.…”

Section: Related Workmentioning

confidence: 99%

Estima

Chatzopoulos

Dragojević

Guerraoui

2017

ACM Trans. Parallel Comput.

View full text Add to dashboard Cite

This article presents estima, an easy-to-use tool for extrapolating the scalability of in-memory applications. estima is designed to perform a simple yet important task: Given the performance of an application on a small machine with a handful of cores, estima extrapolates its scalability to a larger machine with more cores, while requiring minimum input from the user. The key idea underlying estima is the use of stalled cycles (e.g., cycles that the processor spends waiting for missed cache line fetches or busy locks). estima measures stalled cycles on a few cores and extrapolates them to more cores, estimating the amount of waiting in the system. estima can be effectively used to predict the scalability of in-memory applications for bigger execution machines. For instance, using measurements of memcached and SQLite on a desktop machine, we obtain accurate predictions of their scalability on a server. Our extensive evaluation shows the effectiveness of estima on a large number of in-memory benchmarks. INTRODUCTIONCommodity machines nowadays have hundreds of gigabytes of memory. This enables building performance-critical parallel applications, such as databases and key-value stores, that keep their datasets in main memory. This way, applications avoid slow secondary storage and networks, leaving the CPU as the main performance bottleneck [9,12,26,30]. Understanding the performance of these applications proves to be hard, since the number of CPU cores available during the deployment of a parallel application can be significantly higher than that during its development and testing. Applications developed today can be tested on machines with 16 or 24 cores, but in a few years the same applications are likely to be run on machines with 64 or even more cores.

show abstract

Section: Related Workmentioning

confidence: 99%

Estima

Chatzopoulos

Dragojević

Guerraoui

2017

ACM Trans. Parallel Comput.

View full text Add to dashboard Cite

show abstract

“…Usui et al [39] use a simple cost-benefit analysis to choose between locking and transactions. The performance model from [34] focuses on modeling transactional conflict behavior. Unlike ESTIMA, this approach requires heavy instrumentation of the applications in order to collect the statistics of memory accesses.…”

Section: Related Workmentioning

confidence: 99%

Estima

Chatzopoulos

Dragojević

Guerraoui

2016

Proceedings of the 21st ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

View full text Add to dashboard Cite

This paper presents ESTIMA, an easy-to-use tool for extrapolating the scalability of in-memory applications. ESTIMA is designed to perform a simple, yet important task: given the performance of an application on a small machine with a handful of cores, ESTIMA extrapolates its scalability to a larger machine with more cores, while requiring minimum input from the user. The key idea underlying ESTIMA is the use of stalled cycles (e.g. cycles that the processor spends waiting for various events, such as cache misses or waiting on a lock). ESTIMA measures stalled cycles on a few cores and extrapolates them to more cores, estimating the amount of waiting in the system. ESTIMA can be effectively used to predict the scalability of in-memory applications. For instance, using measurements of memcached and SQLite on a desktop machine, we obtain accurate predictions of their scalability on a server. Our extensive evaluation on a large number of in-memory benchmarks shows that ESTIMA has generally low prediction errors.

show abstract

“…Note that the results and ECU values are different, and the fact indicates that we would be better to measure the throughput for predicting the performance of an application running in each instance of Amazon EC2. [9], [25] analyze performance of TM. Heindl and Pokam [25] proposed a framework for performance analysis of STM variants.…”

Section: Performing Dstm Applications In a Public Cloudmentioning

confidence: 99%

“…First, we describe how to execute HyFlow applications in Amazon EC2. We then consider constructing a performance model by adapting a TM performance model proposed by Porter and Witchel [9] to DSTM so that it can take communication costs into account. We discuss some results of experiments using Bank benchmark, which is provided by HyFlow, running in the environment.…”

Section: Introductionmentioning

confidence: 99%

An Experiment on Performing DSTM Applications in a Public Cloud

Yoshino

Aritsugi

2012

2012 41st International Conference on Parallel Processing Workshops

View full text Add to dashboard Cite

Performing distributed software transactional memory (DSTM) applications in a public cloud is investigated in this paper. Transactions are introduced in DSTM for simplifying parallel programming in distributed environments. DSTM is thus a promising alternative to lock-based programming models. Cloud computing attracts attention as a new way for commercial applications and for processing largescale data. However, to our knowledge, there is no study of executing DSTM applications using public clouds. In this paper, we report an experiment on performing DSTM applications in a public cloud. We also try to construct a performance model by adapting a TM performance model to DSTM in order to decide which cloud resources to be chosen in executing DSTM applications. Experimental results show that there are strong and weak points our DSTM model has in deciding machine types and the number of machines for performance.

show abstract

Understanding transactional memory performance

Cited by 13 publications

References 36 publications

Estima

Estima

Estima

An Experiment on Performing DSTM Applications in a Public Cloud

Contact Info

Product

Resources

About