Symbolic Analysis of Concurrency Errors in OpenMP Programs

Computer Aided Verification, 2009

Context-bounded analysis is an attractive approach to verification of concurrent programs. Bounding the number of contexts executed per thread not only reduces the asymptotic complexity, but also the complexity increases gradually from checking a purely sequential program. Lal and Reps [14] provided a method for reducing the context-bounded verification of a concurrent boolean program to the verification of a sequential boolean program, thereby allowing sequential reasoning to be employed for verifying concurrent programs. In this work, we adapt the encoding to work for systems programs written in C with the heap and accompanying low-level operations such as pointer arithmetic and casts. Our approach is completely automatic: we use a verification condition generator and SMT solvers, instead of a boolean model checker, in order to avoid manual extraction of boolean programs and false alarms introduced by the abstraction. We demonstrate the use of field slicing for improving the scalability and (in some cases) coverage of our checking. We evaluate our tool STORM on a set of real-world Windows device drivers, and has discovered a bug that could not be detected by extensive application of previous tools.

IEEE Access

Recently, incorporating more than one programming model into a system designed for high performance computing (HPC) has become a popular solution to implementing parallel systems. Since traditional programming languages, such as C, C++, and Fortran, do not support parallelism, many programmers add one or more programming models to achieve parallelism and accelerate computation efficiently. These models include Open Accelerators (OpenACC) and Open Multi-Processing (OpenMP), which have recently been used with various models, including Message Passing Interface (MPI) and Compute Unified Device Architecture (CUDA). Due to the difficulty of predicting the behavior of threads, runtime errors cannot be predicted. The compiler cannot identify runtime errors such as data races, race conditions, deadlocks, or livelocks. Many studies have been conducted on the development of testing tools to detect runtime errors when using programming models, such as the combinations of OpenACC with MPI models and OpenMP with MPI. Although more applications use OpenACC and OpenMP together, no testing tools have been developed to test these applications to date. This paper presents a testing tool for detecting runtime using a static testing technique. This tool can detect actual and potential runtime errors during the integration of the OpenACC and OpenMP models into systems developed in C++. This tool implement error dependency graphs, which are proposed in this paper. Additionally, a dependency graph of the errors is provided, along with a classification of runtime errors that result from combining the two programming models mentioned earlier.

International Journal of Parallel Programming, 2013

Shared-memory based parallel programming with OpenMP and Posixthread APIs becomes more common to fully take advantage of multiprocessor computing environments. One of the critical risks in multithreaded programming is data races that are hard to debug and greatly damaging to parallel applications if they are uncaught. Although ample effort has been made in building specialized data race detection techniques, the state of art tools such as the Intel Thread Checker still have various functionality and performance problems. In this paper, we present a Versatile On-thefly Race Detection (VORD) tool that provides an agile, efficient, and scalable race detection environment for various parallel programming models. VORD can automatically construct an empirically optimal set of race engines by utilizing classification and adaptation mechanisms. A Race-Detection Classification (RDC) table is created to categorize adequate engines in the aspect of labeling, detecting, and filtering. An Engine Code Property Selector (ECPS) uses the RDC table to adapt optimal engines for the given target programming models. In addition to RDC and ECPS, we have also implemented an OpenMP parser and a source instrumentor. The functionality

Lecture Notes in Computer Science, 2016

We present a new set of tools for the language-centric performance analysis and debugging of OpenMP programs that allows programmers to relate dynamic information from parallel execution to OpenMP constructs. Users can visualize execution traces, examine aggregate metrics on parallel loops and tasks, such as load imbalance or synchronization overhead, and obtain detailed information on specific events, such as the partitioning of a loop's iteration space, its distribution to workers according to the scheduling policy and fine-grain synchronization. Our work is based on the Aftermath performance analysis tool and a ready-to-use, instrumented version of the LLVM/clang OpenMP runtime with negligible overhead for tracing. By analyzing the performance of the MG application of the NPB suite, we show that language-centric performance analysis in general and our tools in particular can help improve the performance of large-scale OpenMP applications significantly.

2016

It is a well-known that data races in implicit threading applications, such as OpenMP programs, are the most notorious class of concurrency bugs, because they may lead to unpredictable results of the program. The main drawback of on-the-fly data race detection techniques is the heavy additional overheads for analyzing every memory operations and thread operations, such as load, store, fork, and join. It is important to reduce the additional overheads for debugging the concurrency bug. This paper presents an empirical comparison of two filtering techniques which are applied to parallel programs with OpenMP directives for dynamic data race detection.

2014

por aceitarem que trabalhasse com eles na minha tese e a todo o grupo Multicore por me ter acolhido e ajudado quando necessitei. Não posso deixar de agradecer também ao meu orientador da Faculdade de Ciências Dr. Dimitris Monstrous, que me ajudou sempre que necessitei e que foi essencial para a minha deslocação para Londres. Muito obrigada sem vocês nada disto teria sido possível. Também quero agradecerà minha família por toda a forca e motivação que me deram. Aos meus pais, Maria e José da Silva, por todo o carinho, força e motivação nestes meses tao intensos de trabalho; a minha irmã, Patrícia que sempre me motivou e animou quando estava mais desanimada e ao meu irmão, Theo que apesar de pequeno esteve sempre ao meu lado quando precisei de uma gargalhada. Obrigada por tudo. Porúltimo, e não menos importante, quero agradecer ao meu namorado, Fábio, que apesar de estar longe sempre me deu muita força mesmo nos momentos em que estava mais stressada e rabugenta; e também quero agradecer aos meus colegas por estes anos todos de faculdade e a duas colegas em especial por me darem muita força quando precisei neste ultimo ano. Obrigada a todos. Muito obrigada a todos que me acompanharam e me deram força nesta etapa da minha vida. iii Á minha família. Resumo Com o uso de linguagens paralelas nascem alguns problemas como data race , uma argura para os programadores durante o seu trabalho. Data race são quando dois ou mais threads em umúnico processo acedem a mesma posição de memória ao mesmo tempo, sendo que um ou mais acessos são para a escrita/leitura, sem sincronização adequada. A nossa ferramenta (SVDR: Static Verification of Data race in OpenMP) foi desenvolvida com o propósito de verificar a ausência de data races em OpenMP (uma linguagem de programação paralela). Empregamos a verificação estática com intenção de encontrar data races sem necessariamente executar o programa, possibilitando dessa forma ao utilizador, encontrar o problema antes de receber o resultado do mesmo. Para melhor compreensão de como funciona a ferramenta,é pertinente apresentar de forma concisa, as linguagens de programação aqui abordadas: OpenMP e Boogie. OpenMP (Aberto Multi-Processing) 2.1é um API usado na programação (C, Fortran, C++) em multiprocessadores com memória partilhada em multiplataformas e funciona em quase todos Sistemas operativos e arquitecturas. Boogie [2, 17]é uma linguagem de verificação intermédia, projectada como uma camada sobre a qualé possível construir outros verificadores para outras línguaguens. Uma vez aclaradas as definições,é mais fácil contextualizar e compreender como a ferramenta funciona. Como já foi referido previamente, o objectivo da ferramentaé traduzir código OpenMP em código Boogie com o propósito de verificar se existe ou não data race no ficheiro de entrada. Assim sendo,é possível aferir que existem duas etapas para o processo de verificação: a primeiraé a tradução de OpenMP em Boogie, e a segundaé a verificação de data race no ficheiro de entrada-queé do tipo bpl, ou seja, código Boogie. A primeira fase nãoé tratado pela ferramenta, de modo queé necessária uma tradução manual do código OpenMP para o código Boogie. A segunda faseé a ferramenta que foi implementada, SVDR. O SVDR funciona da seguinte forma: recebe como entrada um ficheiro bpl (um OpenMP traduzido em código Boogie) e retorna como resultado um ficheiro bpl com o algoritmo necessário para verificar se há ou não um data race no ficheiro de entrada. O algoritmo criadoé baseado no uso de nao determinismo, invariantes e assertions. Este funciona do seguinte modo, para verificar a existência de data race num ciclo While, vii

2004

A program analysis tool can play an important role in helping users understand and improve OpenMP codes. Dragon is a robust interactive program analysis tool based on the Open64 compiler, an open source OpenMP, C/C++/Fortran77/90 compiler for Intel Itanium systems. We developed the Dragon tool on top of Open64 to exploit its powerful analyses in order to provide static as well as dynamic (feedback-based) information which can be used to develop or optimize OpenMP codes. Dragon enables users to visualize and print essential program structures and obtain runtime information on their applications. Current features include static/dynamic call graphs and control flow graphs, data dependence analysis and interprocedural array region summaries, that help understand procedure side effects within parallel loops. On-going work extends Dragon to display data access patterns at runtime, and provide support for runtime instrumentation and optimizations.

2015

Lecture Notes in Computer Science, 2020

Task-based parallel programming models based on compiler directives have proved their effectiveness at describing parallelism in High-Performance Computing (HPC) applications. Recent studies show that cutting-edge Real-Time applications, such as those for unmanned vehicles, can successfully exploit these models. In this scenario, OpenMP is a de facto standard for HPC, and is being studied for Real-Time systems due to its time-predictability and delimited functional safety. However, changes in OpenMP take time to be standardized because it sweeps along a large community. OmpSs, instead, is a task-based model for fast-prototyping that has been a forerunner of OpenMP since its inception. OmpSs-2, its successor, aims at the same goal, and defines several features that can be introduced in future versions of OpenMP. This work targets compiler-based optimizations to enhance the programmability and performance of OmpSs-2. Regarding the former, we present an algorithm to determine the data-sharing attributes of OmpSs-2 tasks. Regarding the latter, we introduce a new algorithm to automatically release OmpSs-2 task dependencies before a task has completed. This work evaluates both algorithms in a set of well-known benchmarks, and discusses their applicability to the current and future specifications of OpenMP.

2021 20th International Symposium on Parallel and Distributed Computing (ISPDC), 2021

Compilers are generally not aware of the semantics of library-based parallel programming models such as MPI and OpenSHMEM, and hence are unable to detect programming errors related to their use. To alleviate this issue, we developed a custom static checker for OpenSHMEM programs based on LLVM's Clang Static Analyzer framework (CSA). We leverage the Symbolic Execution engine of the core Static Analyzer framework and its path-sensitive analysis to check for bugs on all OpenSHMEM program paths. We have identified common programming mistakes in OpenSHMEM programs that are detectable at compile-time and provided checks for them in the analyzer. They cover: utilization of the right type of memory (private vs. symmetric memory); safe/synchronized access to program data in the presence of asynchronous, one-sided communication; and double-free of memories allocated using OpenSHMEM memory allocation routines. Our experimental analysis showed that the static checker successfully detects bugs in OpenSHMEM code.