A multi-GPU algorithm for large-scale neuronal networks

Large-scale simulations of parts of the brain using detailed neuronal models to improve our understanding of brain functions are becoming a reality with the usage of supercomputers and large clusters. However, the high acquisition and maintenance cost of these computers, including the physical space, air conditioning, and electrical power, limits the number of simulations of this kind that scientists can perform. Modern commodity graphical cards, based on the CUDA platform, contain graphical processing units (GPUs) composed of hundreds of processors that can simultaneously execute thousands of threads and thus constitute a low-cost solution for many high-performance computing applications. In this work, we present a CUDA algorithm that enables the execution, on multiple GPUs, of simulations of large-scale networks composed of biologically realistic Hodgkin-Huxley neurons. The algorithm represents each neuron as a CUDA thread, which solves the set of coupled differential equations that model each neuron. Communication among neurons located in different GPUs is coordinated by the CPU. We obtained speedups of 40 for the simulation of 200k neurons that received random external input and speedups of 9 for a network with 200k neurons and 20M neuronal connections, in a single computer with two graphic boards with two GPUs each, when compared with a modern quad-core CPU. Copyright (C) 2010 John Wiley & Sons, Ltd.

A Resource Centric Approach For Advancing Collaboration Through Hydrologic Data And Model Sharing

HydroShare is an online, collaborative system being developed for open sharing of hydrologic data and models. The goal of HydroShare is to enable scientists to easily discover and access hydrologic data and models, retrieve them to their desktop or perform analyses in a distributed computing environment that may include grid, cloud or high performance computing model instances as necessary. Scientists may also publish outcomes (data, results or models) into HydroShare, using the system as a collaboration platform for sharing data, models and analyses. HydroShare is expanding the data sharing capability of the CUAHSI Hydrologic Information System by broadening the classes of data accommodated, creating new capability to share models and model components, and taking advantage of emerging social media functionality to enhance information about and collaboration around hydrologic data and models. One of the fundamental concepts in HydroShare is that of a Resource. All content is represented using a Resource Data Model that separates system and science metadata and has elements common to all resources as well as elements specific to the types of resources HydroShare will support. These will include different data types used in the hydrology community and models and workflows that require metadata on execution functionality. The HydroShare web interface and social media functions are being developed using the Drupal content management system. A geospatial visualization and analysis component enables searching, visualizing, and analyzing geographic datasets. The integrated Rule-Oriented Data System (iRODS) is being used to manage federated data content and perform rule-based background actions on data and model resources, including parsing to generate metadata catalog information and the execution of models and workflows. This presentation will introduce the HydroShare functionality developed to date, describe key elements of the Resource Data Model and outline the roadmap for future development.

A simple mechanism to deal with sequential code in dataflow architectures

The aim of this work is to propose a simple and efficient mechanism to deal with the problem of executing sequential code in a pure dataflow machine. Our results is obtained with a simulator of Wolf [4] architecture. The implemented mechanism improved the architecture performance when executing sequential code and we expect that this improvement could be better if we use some heuristics to deal with some special groups of instructions such as branch operations. Further research will show us if this is true.

An environment for knowledge discovery in biology

This paper describes a data mining environment for knowledge discovery in bioinformatics applications. The system has a generic kernel that implements the mining functions to be applied to input primary databases, with a warehouse architecture, of biomedical information. Both supervised and unsupervised classification can be implemented within the kernel and applied to data extracted from the primary database, with the results being suitably stored in a complex object database for knowledge discovery. The kernel also includes a specific high-performance library that allows designing and applying the mining functions in parallel machines. The experimental results obtained by the application of the kernel functions are reported. © 2003 Elsevier Ltd. All rights reserved.

Scheduler simulation using iSPD, an iconic-based computer grid simulator

Increased accessibility to high-performance computing resources has created a demand for user support through performance evaluation tools like the iSPD (iconic Simulator for Parallel and Distributed systems), a simulator based on iconic modelling for distributed environments such as computer grids. It was developed to make it easier for general users to create their grid models, including allocation and scheduling algorithms. This paper describes how schedulers are managed by iSPD and how users can easily adopt the scheduling policy that improves the system being simulated. A thorough description of iSPD is given, detailing its scheduler manager. Some comparisons between iSPD and Simgrid simulations, including runs of the simulated environment in a real cluster, are also presented. © 2012 IEEE.

iSPD: An iconic-based modeling simulator for distributed grids

Simulation of large and complex systems, such as computing grids, is a difficult task. Current simulators, despite providing accurate results, are significantly hard to use. They usually demand a strong knowledge of programming, what is not a standard pattern in today's users of grids and high performance computing. The need for computer expertise prevents these users from simulating how the environment will respond to their applications, what may imply in large loss of efficiency, wasting precious computational resources. In this paper we introduce iSPD, iconic Simulator of Parallel and Distributed Systems, which is a simulator where grid models are produced through an iconic interface. We describe the simulator and its intermediate model languages. Results presented here provide an insight in its easy-of-use and accuracy.

Técnicas de paralelização em GPGPU aplicadas em algoritmo para remoção de ruído multiplicativo

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Geração de algoritmos de escalonamento para simulação de grades computacionais

Pós-graduação em Ciência da Computação - IBILCE

Image Re-Ranking Acceleration on GPUs

Huge image collections are becoming available lately. In this scenario, the use of Content-Based Image Retrieval (CBIR) systems has emerged as a promising approach to support image searches. The objective of CBIR systems is to retrieve the most similar images in a collection, given a query image, by taking into account image visual properties such as texture, color, and shape. In these systems, the effectiveness of the retrieval process depends heavily on the accuracy of ranking approaches. Recently, re-ranking approaches have been proposed to improve the effectiveness of CBIR systems by taking into account the relationships among images. The re-ranking approaches consider the relationships among all images in a given dataset. These approaches typically demands a huge amount of computational power, which hampers its use in practical situations. On the other hand, these methods can be massively parallelized. In this paper, we propose to speedup the computation of the RL-Sim algorithm, a recently proposed image re-ranking approach, by using the computational power of Graphics Processing Units (GPU). GPUs are emerging as relatively inexpensive parallel processors that are becoming available on a wide range of computer systems. We address the image re-ranking performance challenges by proposing a parallel solution designed to fit the computational model of GPUs. We conducted an experimental evaluation considering different implementations and devices. Experimental results demonstrate that significant performance gains can be obtained. Our approach achieves speedups of 7x from serial implementation considering the overall algorithm and up to 36x on its core steps.

LALP: a language to program custom FPGA-based acceleration engines

Field-Programmable Gate Arrays (FPGAs) are becoming increasingly important in embedded and high-performance computing systems. They allow performance levels close to the ones obtained with Application-Specific Integrated Circuits, while still keeping design and implementation flexibility. However, to efficiently program FPGAs, one needs the expertise of hardware developers in order to master hardware description languages (HDLs) such as VHDL or Verilog. Attempts to furnish a high-level compilation flow (e.g., from C programs) still have to address open issues before broader efficient results can be obtained. Bearing in mind an FPGA available resources, it has been developed LALP (Language for Aggressive Loop Pipelining), a novel language to program FPGA-based accelerators, and its compilation framework, including mapping capabilities. The main ideas behind LALP are to provide a higher abstraction level than HDLs, to exploit the intrinsic parallelism of hardware resources, and to allow the programmer to control execution stages whenever the compiler techniques are unable to generate efficient implementations. Those features are particularly useful to implement loop pipelining, a well regarded technique used to accelerate computations in several application domains. This paper describes LALP, and shows how it can be used to achieve high-performance computing solutions.

Endosymbiosis in trypanosomatids: the genomic cooperation between bacterium and host in the synthesis of essential amino acids is heavily influenced by multiple horizontal gene transfers

Background Trypanosomatids of the genera Angomonas and Strigomonas live in a mutualistic association characterized by extensive metabolic cooperation with obligate endosymbiotic Betaproteobacteria. However, the role played by the symbiont has been more guessed by indirect means than evidenced. Symbiont-harboring trypanosomatids, in contrast to their counterparts lacking symbionts, exhibit lower nutritional requirements and are autotrophic for essential amino acids. To evidence the symbiont’s contributions to this autotrophy, entire genomes of symbionts and trypanosomatids with and without symbionts were sequenced here. Results Analyses of the essential amino acid pathways revealed that most biosynthetic routes are in the symbiont genome. By contrast, the host trypanosomatid genome contains fewer genes, about half of which originated from different bacterial groups, perhaps only one of which (ornithine cyclodeaminase, EC:4.3.1.12) derived from the symbiont. Nutritional, enzymatic, and genomic data were jointly analyzed to construct an integrated view of essential amino acid metabolism in symbiont-harboring trypanosomatids. This comprehensive analysis showed perfect concordance among all these data, and revealed that the symbiont contains genes for enzymes that complete essential biosynthetic routes for the host amino acid production, thus explaining the low requirement for these elements in symbiont-harboring trypanosomatids. Phylogenetic analyses show that the cooperation between symbionts and their hosts is complemented by multiple horizontal gene transfers, from bacterial lineages to trypanosomatids, that occurred several times in the course of their evolution. Transfers occur preferentially in parts of the pathways that are missing from other eukaryotes. Conclusion We have herein uncovered the genetic and evolutionary bases of essential amino acid biosynthesis in several trypanosomatids with and without endosymbionts, explaining and complementing decades of experimental results. We uncovered the remarkable plasticity in essential amino acid biosynthesis pathway evolution in these protozoans, demonstrating heavy influence of horizontal gene transfer events, from Bacteria to trypanosomatid nuclei, in the evolution of these pathways.

A parallel community detection algorithm

Complex networks analysis is a very popular topic in computer science. Unfortunately this networks, extracted from different contexts, are usually very large and the analysis may be very complicated: computation of metrics on these structures could be very complex. Among all metrics we analyse the extraction of subnetworks called communities: they are groups of nodes that probably play the same role within the whole structure. Communities extraction is an interesting operation in many different fields (biology, economics,...). In this work we present a parallel community detection algorithm that can operate on networks with huge number of nodes and edges. After an introduction to graph theory and high performance computing, we will explain our design strategies and our implementation. Then, we will show some performance evaluation made on a distributed memory architectures i.e. the supercomputer IBM-BlueGene/Q "Fermi" at the CINECA supercomputing center, Italy, and we will comment our results.

Algoritmi per la scoperta di outlier in sistemi di calcolo ad altissime prestazioni

In questa tesi vengono analizzati gli algoritmi DistributedSolvingSet e LazyDistributedSolvingSet e verranno mostrati dei risultati sperimentali relativi al secondo.