Library-Independent Data Race Detection

Jannesari, Ali; Tichy, Walter F.

doi:10.1109/tpds.2013.209

Cited by 10 publications

(4 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The parallelized code is expected to deliver the same output as its sequential counterpart. To assure the correctness, we use automated unit testing [10,31,32] and sophisticated race detection methods [11,33]. For data races, our library-independent race detection approach [33] is applied to the generated parallel code for finding potential data races.…”

Section: Correctness Analysismentioning

confidence: 99%

“…To assure the correctness, we use automated unit testing [10,31,32] and sophisticated race detection methods [11,33]. For data races, our library-independent race detection approach [33] is applied to the generated parallel code for finding potential data races. Also, automated unit tests based on dynamic and static analyses are generated, which can be used during the parallelization process for finding atomicity violations and verifying the parallel code.…”

Section: Correctness Analysismentioning

confidence: 99%

See 1 more Smart Citation

Advances in Engineering Software for Multicore Systems

Jannesari¹

2018

Dependability Engineering

Self Cite

View full text Add to dashboard Cite

The vast amounts of data to be processed by today's applications demand higher computational power. To meet application requirements and achieve reasonable application performance, it becomes increasingly profitable, or even necessary, to exploit any available hardware parallelism. For both new and legacy applications, successful parallelization is often subject to high cost and price. This chapter proposes a set of methods that employ an optimistic semi-automatic approach, which enables programmers to exploit parallelism on modern hardware architectures. It provides a set of methods, including an LLVM-based tool, to help programmers identify the most promising parallelization targets and understand the key types of parallelism. The approach reduces the manual effort needed for parallelization. A contribution of this work is an efficient profiling method to determine the control and data dependences for performing parallelism discovery or other types of code analysis. Another contribution is a method for detecting code sections where parallel design patterns might be applicable and suggesting relevant code transformations. Our approach efficiently reports detailed runtime data dependences. It accurately identifies opportunities for parallelism and the appropriate type of parallelism to use as task-based or loop-based.

show abstract

Section: Correctness Analysismentioning

confidence: 99%

Section: Correctness Analysismentioning

confidence: 99%

Advances in Engineering Software for Multicore Systems

Jannesari¹

2018

Dependability Engineering

Self Cite

View full text Add to dashboard Cite

show abstract

“…So far we support only parallel programming languages where locking/unlocking primitives have to be written explicitly in the source code. However, programing languages with implicit synchronization can be easily supported by automatically discovering implicit synchronization patterns [81].…”

Section: Modified Parallelization Strategymentioning

confidence: 99%

“…Parallel programs with implicit mutual exclusion and synchronization mechanism can be supported by existing approaches of detecting synchronizations automatically [81]. This is considered as one of the future-work items.…”

Section: Limitationsmentioning

confidence: 99%

Discovery of Potential Parallelism in Sequential Programs

Jannesari

Wolf

2013

2013 42nd International Conference on Parallel Processing

Self Cite

View full text Add to dashboard Cite

In the era of multicore processors, the responsibility for performance gains has been shifted onto software developers. Once improvements of the sequential algorithm have been exhausted, software-managed parallelism is the only option left. However, writing parallel code is still difficult, especially when parallelizing sequential code written by someone else. A key task in this process is the identification of suitable parallelization targets in the source code. Parallelism discovery tools help developers to find such targets automatically. Unfortunately, tools that identify parallelism during compilation are usually conservative due to the lack of runtime information, and tools relying on runtime information primarily suffer from high overhead in terms of both time and memory. This dissertation presents a generic framework for parallelism discovery based on dynamic program analysis, supporting various types of parallelism while incurring practically affordable overhead. The framework contains two main components: an efficient data-dependence profiler and a set of parallelism discovery algorithms based on a language-independent concept called Computational Unit.The data-dependence profiler serves as the foundation of the parallelism discovery framework. Traditional dependence profiling approaches introduce a tremendous amount of time and memory overhead. To lower the overhead, current methods limit their scope to the subset of the dependence information needed for the analysis they have been created for, sacrificing generality and discouraging reuse. In contrast, the profiler shown in this thesis addresses the problem via signature-based memory management and a lock-free parallel design. It produces detailed dependences not only for sequential but also for multi-threaded code without causing prohibitive overhead, allowing it to serve as a generic base for various program analysis techniques.Computational Units (CUs) provide a language-independent foundation for parallelism discovery. CUs are computations that follow the read-compute-write pattern. Unlike other concepts, they are not restricted to predefined language constructs. A program is represented as a CU graph, in which vertexes are CUs and edges are data dependences. This allows parallelism to be detected that spreads across multiple language constructs, taking code refactoring into consideration. The parallelism discovery algorithms cover both loop and task parallelism.Results of our experiments show that 1) the efficient data-dependence profiler has a very competitive average slowdown of around 80× with accuracy higher than 99.6%; 2) the frame- I would also like to thank my master students and student assistants for their coding support.Wolfram Gottschlich implemented the original version of the signature described in this thesis.Tuan Dung Nguyen implemented the lock-free parallel version of the DiscoPoP profiler. MichaelBeaumont tested the memory skipping technique. Daniel Fried implemented the method of characterizing DOALL loops using machin...

show abstract