High Performance Computing Resources Research Articles

Lessons learned from implementing a national infrastructure in Sweden for storage and analysis of next-generation sequencing data.

Analyzing and storing data and results from next-generation sequencing (NGS) experiments is a challenging task, hampered by ever-increasing data volumes and frequent updates of analysis methods and tools. Storage and computation have grown beyond the capacity of personal computers and there is a need for suitable e-infrastructures for processing. Here we describe UPPNEX, an implementation of such an infrastructure, tailored to the needs of data storage and analysis of NGS data in Sweden serving various labs and multiple instruments from the major sequencing technology platforms. UPPNEX comprises resources for high-performance computing, large-scale and high-availability storage, an extensive bioinformatics software suite, up-to-date reference genomes and annotations, a support function with system and application experts as well as a web portal and support ticket system. UPPNEX applications are numerous and diverse, and include whole genome-, de novo- and exome sequencing, targeted resequencing, SNP discovery, RNASeq, and methylation analysis. There are over 300 projects that utilize UPPNEX and include large undertakings such as the sequencing of the flycatcher and Norwegian spruce. We describe the strategic decisions made when investing in hardware, setting up maintenance and support, allocating resources, and illustrate major challenges such as managing data growth. We conclude with summarizing our experiences and observations with UPPNEX to date, providing insights into the successful and less successful decisions made.

A Big Data Approach to Analyzing Market Volatility

Understanding the microstructure of the financial market requires the processing of a vast amount of data related to individual trades, and sometimes even multiple levels of quotes. Analyzing such a large volume of data requires tremendous computing power that is not easily available to financial academics and regulators. Fortunately, public funded High Performance Computing (HPC) power is widely available at the National Laboratories in the US. In this paper we demonstrate that the HPC resource and the techniques for data-intensive sciences can be used to greatly accelerate the computation of an early warning indicator called Volume-synchronized Probability of Informed trading (VPIN). The test data used in this study contains five and a half year's worth of trading data for about 100 most liquid futures contracts, includes about 3 billion trades, and takes 140GB as text files. By using (1) a more efficient file format for storing the trading records, (2) more effective data structures and algorithms, and (3) parallelizing the computations, we are able to explore 16,000 different ways of computing VPIN in less than 20 hours on a 32-core IBM DataPlex machine. Our test demonstrates that a modest computer is sufficient to monitor a vast number of trading activities in real-time -- an ability that could be valuable to regulators.Our test results also confirm that VPIN is a strong predictor of liquidity-induced volatility. With appropriate parameter choices, the false positive rates are about 7% averaged over all the futures contracts in the test data set. More specifically, when VPIN values rise above a threshold (CDF > 0.99), the volatility in the subsequent time windows is higher than the average in 93% of the cases.

HPC in Germany: Infrastructure, Operations and Politics

Abstract The overall scientific challenge for most areas of science is to master the presence of massive parallelism and, with its help, to allow high-end applications to benefit from the high-performance computing (HPC) technology available today. Many key issues need to be addressed, ranging from parallel programming paradigms, via dynamic load distribution, to energy-awareness, and scalability of algorithms and applications, to name just a few. This paper gives an overview of the HPC landscape in Germany. With the Gauss Centre for Supercomputing (GCS) and the Gauss Alliance (GA), the structure of the German HPC centres offers users from all scientific areas access to HPC resources at all potentially required levels of performance. However, the significant contribution of these centres — beyond providing the hardware infrastructure — is to push software technology on all required key issues mentioned above, to make the access to HPC hardware resources efficient. In spite of the reasonable hardware funding of the GCS and GA HPC centres, the critical issue of finding the right “brains” to program the “racks” still continues to remain open.

Computational fluid dynamics (CFD) simulation of liquid aerosol coalescing filters

This work presents a new computational fluid dynamics model, developed to simulate the behaviour of fibrous filters used to treat liquid aerosols. We demonstrate that the solver can resolve the fundamental processes that occur in coalescing filters, such as aerosol droplet capture, coalescence, film break-up, and the motion of coalesced liquid droplets. Initial simulations were conducted on idealised geometries to ensure agreement between the simulations and existing theory, before simulations on more complex geometries were carried out, and finally simulations on realistic filter geometries. It is demonstrated that CFD can provide a viable alternative to whole-scale mist filter testing, particularly if high performance computational resources are available. Additionally, the paper includes a comprehensive treatment of the issues to be overcome when using Lagrangian methods to simulate nano-sized aerosols.

Presentation and analysis of the IPSL and CNRM climate models used in CMIP5

After several years of development, the French IPSL and CNRM-CERFACS modelling groups are presenting to the international scientific community the climate simulations they have prepared as part of the 5th Phase of the Coupled Model Intercomparison Project (CMIP5). The following set of manuscripts provides a detailed description and evaluation of the coupled models and of their individual components. It also presents an analysis of the simulations in terms of internal climate variability, climate response to natural and anthropogenic forcings, physical, chemical and biogeochemical feedbacks, and regional studies. This work benefitted tremendously from the national INSU/LEFE ‘‘MissTerre’’ project (P.I. P. Braconnot and S. Planton) which aims at strengthening collaborations among the French research teams involved in climate modeling. Several workshops and substantial data and expertise exchanges has been possible thanks to this project. This joint Special Issue analyzing simulations from the two models is one of the main results of this collaboration. This work was also supported by other national and international projects cited in the relevant articles. This work was made possible thanks to the High-Performance Computing resources of CCRT and IDRIS made available by GENCI (Grand Equipement National de Calcul Intensif), CEA (Commissariat a l’Energie Atomique et aux Energies Alternatives) and CNRS (Centre National de la Recherche Scientifique) and thanks to Meteo-France computing facility. More information about the IPSL-CM5 and CNRMCM5 climate models can be found at: http://icmc.ipsl.fr/ and at http://www.cnrm.meteo.fr/cmip5/, respectively. We are most grateful to the many people from the IPSL Climate Modelling Centre, CNRM and CERFACS who made possible the production, the preparation and the distribution of the IPSL-CM5 and CNRM-CM5 model results. This special collection was coordinated by Juliette Mignot and Sandrine Bony

CartograTree: connecting tree genomes, phenotypes and environment

Today, researchers spend a tremendous amount of time gathering, formatting, filtering and visualizing data collected from disparate sources. Under the umbrella of forest tree biology, we seek to provide a platform and leverage modern technologies to connect biotic and abiotic data. Our goal is to provide an integrated web-based workspace that connects environmental, genomic and phenotypic data via geo-referenced coordinates. Here, we connect the genomic query web-based workspace, DiversiTree and a novel geographical interface called CartograTree to data housed on the TreeGenes database. To accomplish this goal, we implemented Simple Semantic Web Architecture and Protocol to enable the primary genomics database, TreeGenes, to communicate with semantic web services regardless of platform or back-end technologies. The novelty of CartograTree lies in the interactive workspace that allows for geographical visualization and engagement of high performance computing (HPC) resources. The application provides a unique tool set to facilitate research on the ecology, physiology and evolution of forest tree species. CartograTree can be accessed at: http://dendrome.ucdavis.edu/cartogratree.

A unified framework for the deployment, exposure and access of HPC applications as services in clouds

Clouds have provided on-demand, scalable and affordable High Performance Computing (HPC) resources to discipline (e.g., Biology, Medicine, Chemistry) scientists. However, the steep learning curve of preparing a HPC cloud and deploying HPC applications has hindered many scientists to achieve innovative discoveries for which HPC resources must be relied on. With the world moving to web-based tools, scientists are also seeking more web-based technologies to support their research. Unfortunately, the discipline problems of high-performance computational research are both unique and complex, which make the development of web-based tools for this research difficult. This paper presents our work on developing a unified cloud framework that allows discipline users to easily deploy and expose HPC applications in public clouds as services. To provide a proof of concept, we have implemented the cloud framework prototype by integrating three components: (i) Amazon EC2 public cloud for providing HPC infrastructure, (ii) a HPC service software library for accessing HPC resources, and (iii) the Galaxy web-based platform for exposing and accessing HPC application services. This new approach can reduce the time and money needed to deploy, expose and access discipline HPC applications in clouds.

Pragmatic optimizations for better scientific utilization of large supercomputers

Advances in modeling and algorithms, combined with growth in computing resources, have enabled simulations of multiphysics–multiscale phenomena that can greatly enhance our scientific understanding. However, on currently available high-performance computing (HPC) resources, maximizing the scientific outcome of simulations requires many trade-offs. In this paper we describe our experiences in running simulations of the explosion phase of Type Ia supernovae on the largest available platforms. The simulations use FLASH, a modular, adaptive mesh, parallel simulation code with a wide user base. The simulations use multiple physics components: hydrodynamics, gravity, a sub-grid flame model, a three-stage burning model, and a degenerate equation of state. They also use Lagrangian tracer particles, which are then post-processed to determine the nucleosynthetic yields. We describe the simulation planning process, and the algorithmic optimizations and trade-offs that were found to be necessary. Several of the optimizations and trade-offs were made during the course of the simulations as our understanding of the challenges evolved, or when simulations went into previously unexplored physical regimes. We also briefly outline the anticipated challenges of, and our preparations for, the next-generation computing platforms.

Cloud CPFP: A Shotgun Proteomics Data Analysis Pipeline Using Cloud and High Performance Computing

We have extended the functionality of the Central Proteomics Facilities Pipeline (CPFP) to allow use of remote cloud and high performance computing (HPC) resources for shotgun proteomics data processing. CPFP has been modified to include modular local and remote scheduling for data processing jobs. The pipeline can now be run on a single PC or server, a local cluster, a remote HPC cluster, and/or the Amazon Web Services (AWS) cloud. We provide public images that allow easy deployment of CPFP in its entirety in the AWS cloud. This significantly reduces the effort necessary to use the software, and allows proteomics laboratories to pay for compute time ad hoc, rather than obtaining and maintaining expensive local server clusters. Alternatively the Amazon cloud can be used to increase the throughput of a local installation of CPFP as necessary. We demonstrate that cloud CPFP allows users to process data at higher speed than local installations but with similar cost and lower staff requirements. In addition to the computational improvements, the web interface to CPFP is simplified, and other functionalities are enhanced. The software is under active development at two leading institutions and continues to be released under an open-source license at http://cpfp.sourceforge.net.

A Game-Theoretic Analysis of Grid Job Scheduling

Computational Grid is a well-established platform that gives an assurance to provide a vast range of heterogeneous resources for high performance computing. Efficient and effective resource management and Grid job scheduling are key requirements in order to optimize the use of the resources and to take full advantage from Grid systems. In this paper, we study the job scheduling problem in Computational Grid by using a game-theoretic approach. Grid resources are usually owned by different organizations which may have different and possibly conflicting concerns. Thus it is a crucial objective to analyze potential scenarios where selfish or cooperative behaviors of organizations impact heavily on global Grid efficiency. To this purpose, we formulate a repeated non-cooperative job scheduling game, whose players are Grid sites and whose strategies are scheduling algorithms. We exploit the concept of Nash equilibrium to express a situation in which no player can gain any profit by unilaterally changing its strategy. We extend and complement our previous work by showing whether, under certain circumstances, each investigated strategy is a Nash equilibrium or not. In the negative case we give a counter-example, in the positive case we either give a formal proof or motivate our conjecture by experimental results supported by simulations and exhaustive search.

Real-time Visualisation and Analysis of Tera-scale Datasets

AbstractAs we move ever closer to the Square Kilometre Array era, support for real-time, interactive visualisation and analysis of tera-scale (and beyond) data cubes will be crucial for on-going knowledge discovery. However, the data-on-the-desktop approach to analysis and visualisation that most astronomers are comfortable with will no longer be feasible: tera-scale data volumes exceed the memory and processing capabilities of standard desktop computing environments. Instead, there will be an increasing need for astronomers to utilise remote high performance computing (HPC) resources. In recent years, the graphics processing unit (GPU) has emerged as a credible, low cost option for HPC. A growing number of supercomputing centres are now investing heavily in GPU technologies to provide O(100) Teraflop/s processing. I describe how a GPU-powered computing cluster allows us to overcome the analysis and visualisation challenges of tera-scale data. With a GPU-based architecture, we have moved the bottleneck from processing-limited to bandwidth-limited, achieving exceptional real-time performance for common visualisation and data analysis tasks.

Optimization of Shared High-Performance Reconfigurable Computing Resources

In the field of high-performance computing, systems harboring reconfigurable devices, such as field-programmable gate arrays (FPGAs), are gaining more widespread interest. Such systems range from supercomputers with tightly coupled reconfigurable hardware to clusters with reconfigurable devices at each node. The use of these architectures for scientific computing provides an alternative for computationally demanding problems and has advantages in metrics, such as operating cost/performance and power/performance. However, performance optimization of these systems can be challenging even with knowledge of the system’s characteristics. Our analytic performance model includes parameters representing the reconfigurable hardware, application load imbalance across the nodes, background user load, basic message-passing communication, and processor heterogeneity. In this article, we provide an overview of the analytical model and demonstrate its application for optimization and scheduling of high-performance reconfigurable computing (HPRC) resources. We examine cost functions for minimum runtime and other optimization problems commonly found in shared computing resources. Finally, we discuss additional scheduling issues and other potential applications of the model.

A GPU-Based Parallel Processing Method for Slope Analysis in Geographical Computation

Nowadays, geographical computation presents to be distributed, parallel, and diversification application trend. Influence of problem scale and response speed requirement received more attention. And high performance computational systems, such as “TianHe 1-A”, provides a new generation of hardware supports. In order to make full use of these high performance computational resources, appropriate and efficient parallel algorithms are needed. New parallel computing optimization technique of the geography is proposed in this paper by designing new parallel algorithms for slope analysis. We implement it based on CUDA. Experimental results with random DEM data in uniform distribution validate our methods.

High-performance computing tools for the integrated assessment and modelling of social–ecological systems

Integrated spatio-temporal assessment and modelling of complex social–ecological systems is required to address global environmental challenges. However, the computational demands of this modelling are unlikely to be met by traditional Geographic Information System (GIS) tools anytime soon. I evaluated the potential of a range of high-performance computing (HPC) hardware and software tools to overcome these computational barriers. Performance advantages were quantified using a synthetic model. Four tests were compared, using: a) an Arc Macro Language (AML) GIS script on a single central processing unit (CPU); b) Python/NumPy on 1–256 CPU cores; c) Python/NumPy on 1–64 graphics processing units (GPUs) with high-level PyCUDA abstraction (GPUArray); and d) Python/NumPy on 1–64 GPUs with low-level PyCUDA abstraction (ElementwiseKernel). The GIS implementation effectively took 15.5 weeks to run. Python/NumPy on a single CPU core led to a speed-up of 59× compared to the GIS. On a single GPU, speed-ups of 1473× were achieved using GPUArray and 4881× using ElementwiseKernel. Parallel processing led to further performance enhancements. At best, the ElementwiseKernel module in parallel over 64 GPUs achieved a speed-up of 63,643×. Open source tools such as Python applied across a spectrum of HPC resources offer transformational and accessible performance improvements for integrated assessment and modelling. By reducing the computational barrier, HPC can lead to a step change in modelling sophistication, including the better representation of uncertainty, and perhaps even new modelling paradigms. However, migration to new hardware and software environments also has significant costs. Costs and benefits of HPC are discussed and code tools are provided to help others migrate to HPC and transform our ability to address global challenges through integrated assessment and modelling.

Big Computers for Little Engineers

The author explores the potential for high-performance computing (HPC) to save time and money in new product innovation by U.S. small manufacturers. He describes the U.S. National Digital Engineering and Manufacturing Consortium (NDEMC) program to improve small manufacturers' access to HPC resources such as software and expertise, and gives the example of Jeco Plastic Products, which used supercomputer simulations of the performance of their plastic pallet designs instead of prototype testing.

HiTempo: a platform for time-series analysis of remote-sensing satellite data in a high-performance computing environment

Course resolution earth observation satellites offer large data sets with daily observations at global scales. These data sets represent a rich resource that, because of the high acquisition rate, allows the application of time-series analysis methods. To research the application of these time-series analysis methods to large data sets, it is necessary to turn to high-performance computing (HPC) resources and software designs. This article presents an overview of the development of the HiTempo platform, which was designed to facilitate research into time-series analysis of hyper-temporal sequences of satellite image data. The platform is designed to facilitate the exhaustive evaluation and comparison of algorithms, while ensuring that experiments are reproducible. Early results obtained using applications built within the platform are presented. A sample model-based change detection algorithm based on the extended Kalman filter has been shown to achieve a 97% detection success rate on simulated data sets constructed from MODIS time series. This algorithm has also been parallelized to illustrate that an entire sequence of MODIS tiles (415 tiles over 9 years) can be processed in under 19 minutes using 32 processors.

Performance analysis of Intel multiprocessors using astrophysics simulations

SUMMARYThis paper provides a performance evaluation and investigation of the astrophysics code FLASH for a variety of Intel multiprocessors. This work was performed at the NASA Center for Computational Sciences (NCCS) on behalf of the Carnegie Institution of Washington (CIW) as a study preliminary to the acquisition of a high‐performance computing (HPC) system at the CIW and for the NCCS itself to measure the relative performance of a recently acquired Intel Nehalem‐based system against previously installed multicore HPC resources. A brief overview of computer performance evaluation is provided, followed by a description of the systems under test, a description of the FLASH test problem, and the test results. Additionally, the paper characterizes some of the effects of load imbalance imposed by adaptive mesh refinement. Copyright © 2012 John Wiley & Sons, Ltd.

Hybrid Cloud Computing Platform for Hazardous Chemicals Releases Monitoring and Forecasting

While hazardous chemicals pollutions occurring, monitoring and forecasting are very important for emergency planning and evacuation. To implement the object, huge amounts of wireless sensors may be distributed in large area. The data gathered by the sensors are sent back to data centers for preprocessing and analysis, which need massive computing utility. Therefore a high effective computing paradigm is essential for the hazardous chemicals releases monitoring and forecasting. This paper brought out a hybrid cloud computing platform to solve the problem. It is an Infrastructure as a Service (IaaS) platform, which provides High Performance Computing (HPC) resources, virtualized computing resources, storage resources for hazardous chemicals releases data processing and analysis. Especially we designed a scheduling algorithm and QoS policy to assure efficiency of the platform. At last the platform capability and performance were demonstrated by application scenarios.

Application Scenarios Using Serpens Suite for Kepler Scientific Workflow System

This paper presents the overview of exploitation scenarios making use of the Serpens suite for the Kepler workflow orchestration system. The proposed framework provides researchers with an easy-to-use, workflow-based environment for scientific computations. It allows execution of various applications coming from different disciplines, in various distributed computational environments using a user-friendly interface. This research has been driven initially by Nuclear Fusion applications’ requirements, where the leading idea was to enhance the modeling capabilities for ITER sized plasma research by providing access to High Performance Computing resources. Several usage scenarios are being presented with an example of applications rom the field of Nuclear Fusion, Astrophysics and Computational Chemistry.

Are There Clouds in Our Blue Sky Research Programs?

Guest editorial As our industry pushes the frontiers of hydrocarbon recovery, complexity and risk management are increasingly pervasive dimensions of the landscape. High-performance computing (HPC) is providing new ways to address complexity and risk by opening more workflows to the “real time” world of operations. Cray supercomputers—named for inventor Seymour Cray—were introduced in the 1960s and became synonymous with this new breed of computer. Supercomputer performance is measured in floating point operations per second (FLOPS), which measures the number of calculations or instructions performed in a given time. This is generally used with an International System of Units (SI) prefix such as mega-, giga-, tera-, or peta- to describe the machine’s power or class. Today’s supercomputers are “petascale” machines, capable of processing one quadrillion (1015) FLOPS. However, they are the size of several rooms and require megawatts of power to operate, making them impractical and costly for all but the most esoteric and specialized problems. Computer clusters have become a more common way to build this processing power. By combining high-speed local networks and associated software, or “middleware,” massive parallel systems can be built that rival supercomputing power. Cloud computing is the latest extension of this concept in which computer resources, processing, and memory can be temporarily built into the required computing framework. Add in new job scheduling software, advanced virtualization techniques, and dynamic architecture configurations and you have the rapidly expanding high-performance computing discipline. Our industry is turning to high performance and cloud computing as a cost-effective way to build on-demand supercomputer capability for our most complex or intensive problems. At Baker Hughes, a three-tier approach has been designed to unlock this processing power for our research and development groups. The Dedicated Tier is an array of high-performance central processing units and high-memory servers that can provide massive total compute power joined into clusters with fast interconnects. This configuration is used mainly for applications that require dedicated, uninterrupted access to hardware and benefit from parallelization in a traditional, high-performance computing environment. The available performance capacity is often in the tens of teraFLOPS. The Opportunistic Tier is basically “cycle scavenging,” or the use of idle assets such as desktop computers and unused capacity on servers. Using Wake on LAN (WOL) technology, these computers can be powered on or off as needed. This approach is suitable for short-run applications with small footprints. It is useful where the synchronization among parallel jobs can be loosely coupled as well as when a user is far away from traditional HPC resources, but close to underused powerful desktops and servers. The available power scales linearly with the number of personal computers and nondedicated servers; it is not unusual to see power capabilities in the hundreds of teraFLOPS or more at a Fortune 250 company.