Jean Luca Bez

Lawrence Berkeley National Laboratory | LBL · Computational Research Division (CRD)

We present a comprehensive survey on parallel I/O in the high-performance computing (HPC) context. This is an important field for HPC because of the historic gap between processing power and storage latency, which causes application performance to be impaired when accessing or generating large amounts of data. As the available processing power and...

In this paper, we propose an approach to adapt the I/O forwarding layer of HPC systems to applications’ access patterns. I/O optimization techniques can improve performance for the access patterns they were designed to target, but they often decrease performance for others. Furthermore, these techniques usually depend on the precise tune of their p...

The existing parallel I/O stack is complex and difficult to tune due to the interdependencies among multiple factors that impact the performance of data movement between storage and compute systems. When performance is slower than expected, end-users, developers, and system administrators rely on I/O profiling and tracing information to pinpoint th...

Using parallel file systems efficiently is a tricky problem due to inter-dependencies among multiple layers of I/O software, including high-level I/O libraries (HDF5, netCDF, etc.), MPI-IO, POSIX, and file systems (GPFS, Lustre, etc.). Profiling tools such as Darshan collect traces to help understand the I/O performance behavior. However, there are...

I/O operations are a bottleneck for numerous applications, so optimizing the performance of these operations is of paramount importance. Many techniques explore and apply optimizations to different layers of the I/O stack to improve performance. The difficulty that arises is that the workload changes constantly. So detecting access patterns correct...

Parallel I/O is a critical technique for moving data between compute and storage subsystems of supercomputing systems. With massive amounts of data being produced or consumed by compute nodes, high performant parallel I/O is essential. I/O benchmarks play an important role in this process, however, there is a scarcity of I/O benchmarks that are rep...

Input and output (I/O) operations are a bottleneck for an increasing number of applications in High-Performance Computing (HPC) platforms. Furthermore, it has the potential of critically impacting performance on the next generation of supercomputers. I/O optimization techniques can provide improvements for specific system configurations and applica...

Os métodos de avaliação que mais se destacam são aqueles que apresentam um panorama preciso da evolução das múltiplas habilidades dos estudantes. Dessa forma, apresentamos o modelo TriMElo que tem por objetivo estimar simultaneamente a habilidade geral e as sub-habilidades dos estudantes de acordo com as habilidades envolvidas nos problemas. O mode...

É crescente o número de plataformas online que disponibilizam exercícios de programação, onde os estudantes submetem a resolução destes exercícios e recebem um feedback automático do sistema, sem intervenção humana. Esses ambientes permitem o registro de muitos aspectos das submissões e, dessa forma, os modelos de avaliação educacional podem ser ut...

Nas aplicações de HPC operações de E/S são gargalos devido a diferença entre velocidade de processamento e de acesso a um dado. Assim, caracterizar as operações pode auxiliar na busca de desempenho. Deste modo, este trabalho propõe uma abordagem com aprendizado não supervisionado para a caracterização de E/S. Utilizando dados do supercomputador Int...

Este estudo busca identificar os tamanhos de requisições de entrada e saída mais comuns utilizados por aplicações HPC em ambientes de grande escala. Para isso, utilizamos dados de um ano inteiro de caracterização com a ferramenta Darshan no supercomputador Intrepid Blue Gene/P. Ao identificar os diferentes padrões de acesso e os tamanhos de requisi...

In this article, we study the I/O performance of the Santos Dumont supercomputer, since the gap between processing and data access speeds causes many applications to spend a large portion of their execution on I/O operations. For a large-scale expensive supercomputer, it is essential to ensure applications achieve the best I/O performance to promot...

A personalização em ambientes educacionais demanda modelos próprios. Umfator importante na mediação é apresentar objetos de aprendizagem que catalisem o aprendizado, respeitando as necessidades educacionais do estudante. Neste trabalho apresentamos um modelo de mediação amparado em um modelo pedagógico cognitivo. A ideia central do modelo é recomen...

Métodos de avaliação da proficiência dos estudantes vem ganhando destaque nos últimos anos. Existe um número crescente de cursos online e plataformas que disponibilizam repositórios de questões ou exercícios onde os métodos de avaliação ocorrem de forma automática. Esse trabalho faz uma análise dos dados gerados através de dois modelos que tem por...

In this paper, we propose a pattern matching approach for server-side access pattern detection for the HPC I/O stack. More specifically, our proposal concerns file-level accesses, such as the ones made to I/O libraries, I/O nodes, and the parallel file system servers. The goal of this detection is to allow the system to adapt to the current workloa...

In this research, we are working towards the dynamic reconfiguration of the I/O forwarding layer, adapting it to the needs of the set of applications running in the supercomputer. The I/O forwarding layer is placed between processing nodes and the parallel file system, and its I/O nodes serve to decrease the concurrency in the access to the shared...

This paper presents an energy efficiency and I/O performance analysis of low‐power architectures when compared to conventional architectures, with the goal of studying the viability of using them as storage servers. Our results show that despite the fact the power demand of the storage device amounts for a small fraction of the power demand of the...

This paper presents a new forum for the URI Online Judge website, aimed at solving problems observed during its first version. With a consistent, more intuitive and better-organized interface, we seek to strengthen the teaching-learning process by encouraging the exchange of knowledge and the sharing of experiences in programming.

Este artigo apresenta uma ferramenta online de execução de scripts SQL, integrada com a ferramenta URI Online Judge. O URI SQL é uma ferramenta em fase final de desenvolvimento que apoiará alunos e professores nas disciplinas de Banco de Dados, apresentando um ambiente visual para submissão e testes de execução de scripts em linguagem de consulta.

As large-scale parallel platforms are deployed to comply with the increasing performance requirements of scientific applications, a new concern is getting the attention of the HPC community: the power consumption. In this paper, we aim at evaluating the viability of using low-power architectures as file systems servers in HPC environments, since pr...

In High Performance Computing (HPC) environments, scientific applications rely on Parallel File Systems (PFS) to obtain Input/Output (I/O) performance especially when handling large amounts of data. However, I/O is still a bottleneck for an increasing number of applications, due to the historical gap between processing and data access speed. To all...

Este artigo apresenta uma análise de desempenho e eficiência energética de operações de E/S em processadores de baixo consumo quando comparados a arquiteturas convencionais. O objetivo é analisar a viabilidade da utilização destes dispositivos na implementação de sistemas de arquivos para HPC. Os resultados mostraram que o uso do MPSoC levou a uma e...

We analyze the feasibility and energy efficiency of using an unconventional cluster of low-power Advanced RISC Machines processors to execute two scientific parallel applications. For this purpose, we have selected two applications that present high computational and communication cost: the Ondes3D that simulates geophysical events, and the all-pai...

URI Online Judge is a fast growing web platform designed for students and professors. Our problem archive contains more than a thousand challenges that can be solved in five different programming languages. The categorization of problems help our users to focus on specific programming topics and levels of complexity. Social network concepts were in...

Available at http://hal.univ-grenoble-alpes.fr/hal-01591755

In this paper, we aimed at analyzing the feasibility and energy efficiency when using an unconventional cluster of ARM processors to execute a scientific application. For this purpose, we used Ondes3D to simulate geophysical events. We present a comparison using different compilation flags and distinct values for the processors frequency. It was po...

ARM processors are beginning to gain attention from the HPC community due to its performance and energy efficiency characteristics. When developing HPC applications for such testbeds developers assume that the computation resources available are homogeneous. However, we observed some anomalies when executing a relatively simple HPC application (an...

This work presents the parallel storage device profiling tool SeRRa. Our tool obtains the sequential to random throughput ratio for reads and writes of different sizes on storage devices. In order to provide this information efficiently, SeRRa employs benchmarks to obtain the values for only a subset of the parameter space and estimates the remaini...

This work presents the parallel storage device profiling tool SeRRa. Our tool obtains the sequential to random throughput ratio for reads and writes of different sizes on storage devices. In order to provide this information efficiently, SeRRa employs benchmarks to obtain the values for only a subset of the parameter space and estimates the remaini...

O propósito deste artigo é apresentar a última versão do módulo Academic, uma ferramenta integrada ao portal URI Online Judge. O Academic, ambiente de gerenciamento de trabalhos e listas de exercícios, foi criado em 2013 com o objetivo de facilitar as atividades didáticas de professores e coaches de algoritmos e linguagens de programação. Por ser d...

The URI Online Judge Academic is an online tool that assists professors in programming classes and motivates students to practice more and to go beyond the theoretical base learned in class and, as a result, helps them sharpen their logical, algorithmically and programming skills. The Academic module enables professors to manage disciplines and lis...

There is an increasing demand for tools that help the learning process in all areas of knowledge. The students have more and more difficulties with the traditional classes based on the exposition of the contents in the blackboard or slides. We need new methods and tools to better explore the new technologies available. The URI Online Judge is such...

URI Online Judge is an online tool that has a great potential for enhancing the teaching and learning process if it is aggregated with Algorithms classes. The tool allows problem correction in real time, interaction between users and flexibility in terms of the choice of the programming language. The professor does not have to deal with the complex...