DSpace UNL

RUN >
Faculdade de Ciências e Tecnologia (FCT) >
FCT Departamentos >
FCT: Departamento de Informática >
FCT: DI - Dissertações de Mestrado >

Please use this identifier to cite or link to this item: http://hdl.handle.net/10362/5128

Title: Static detection of anomalies in transactional memory programs
Authors: Teixeira, Bruno Cavaca
Advisor: Lourenço, João
Keywords: Transactional memory
Concurrent programming
Concurrency anomalies
Program testing
Program validation
Static analysis
Issue Date: 2010
Publisher: Faculdade de Ciências e Tecnologia
Abstract: Transactional Memory (TM) is an approach to concurrent programming based on the transactional semantics borrowed from database systems. In this paradigm, a transaction is a sequence of actions that may execute in a single logical instant, as though it was the only one being executed at that moment. Unlike concurrent systems based in locks, TM does not enforce that a single thread is performing the guarded operations. Instead, like in database systems, transactions execute concurrently, and the effects of a transaction are undone in case of a conflict, as though it never happened. The advantages of TM are an easier and less error-prone programming model, and a potential increase in scalability and performance. In spite of these advantages, TM is still a young and immature technology, and has still to become an established programming model. It still lacks the paraphernalia of tools and standards which we have come to expect from a widely used programming paradigm. Testing and analysis techniques and algorithms for TM programs are also just starting to be addressed by the scientific community, making this a leading research work is many of these aspects. This work is aimed at statically identifying possible runtime anomalies in TMprograms. We addressed both low-level dataraces in TM programs, as well as high-level anomalies resulting from incorrect splitting of transactions. We have defined and implemented an approach to detect low-level dataraces in TM programs by converting all the memory transactions into monitor protected critical regions, synchronized on a newly generated global lock. To validate the approach, we have applied our tool to a set of tests, adapted from the literature, that contain well documented errors. We have also defined and implemented a new approach to static detection of high-level concurrency anomalies in TM programs. This new approach works by conservatively tracing transactions, and matching the interference between each consecutive pair of transactions against a set of defined anomaly patterns. Once again, the approach was validated with well documented tests adapted from the literature.
Description: Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para a obtenção do Grau de Mestre em Engenharia Informática
URI: http://hdl.handle.net/10362/5128
Appears in Collections:FCT: DI - Dissertações de Mestrado

Files in This Item:

File Description SizeFormat
Teixeira_2010.pdf1.44 MBAdobe PDFView/Open
Statistics
View Statistics
FacebookTwitterDeliciousLinkedInDiggGoogle BookmarksMySpaceOrkut
Formato BibTex mendeley Endnote Logotipo do DeGóis 

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Universidade Nova de Lisboa  - Statistics  - Feedback
Estamos no RCAAP Governo Português separator Ministério da Educação e Ciência   Fundação para a Ciência e a Tecnologia

Financiado por:

POS_C UE