Paderborn Center for Parallel Computing

Paderborn, Germany

Paderborn Center for Parallel Computing

Paderborn, Germany
Time filter
Source Type

Meister D.,Paderborn Center for Parallel Computing | Brinkmann A.,Paderborn Center for Parallel Computing
2010 IEEE 26th Symposium on Mass Storage Systems and Technologies, MSST2010 | Year: 2010

Data deduplication systems discover and remove redundancies between data blocks. The search for redundant data blocks is often based on hashing the content of a block and comparing the resulting hash value with already stored entries inside an index. The limited random IO performance of hard disks limits the overall throughput of such systems, if the index does not fit into main memory. This paper presents the architecture of the dedupv1 deduplication system that uses solid-state drives (SSDs) to improve its throughput compared to disk-based systems. dedupv1 is designed to use the sweet spots of SSD technology (random reads and sequential operations), while avoiding random writes inside the data path. This is achieved by using a hybrid deduplication design. It is an inline deduplication system as it performs chunking and fingerprinting online and only stores new data, but it is able to delay much of the processing as well as IO operations. ©2010 IEEE.

Calcavecchia F.,Johannes Gutenberg University Mainz | Calcavecchia F.,CNRS Physics and Models in Condensed Media Laboratory | Kuhne T.D.,Johannes Gutenberg University Mainz | Kuhne T.D.,Paderborn Center for Parallel Computing | Kuhne T.D.,University of Paderborn
EPL | Year: 2015

We demonstrate that extending the shadow wave function to fermionic systems facilitates to accurately calculate strongly correlated multi-reference systems such as the stretched H2 molecule. This development considerably extends the scope of electronic-structure calculations and enables to efficiently recover the static correlation energy using just a single Slater determinant. Copyright © EPLA, 2015.

Keller M.,Paderborn Center for Parallel Computing | Meister D.,Paderborn Center for Parallel Computing | Brinkmann A.,Paderborn Center for Parallel Computing | Terboven C.,RWTH Aachen | Bischof C.,RWTH Aachen
Proceedings - 37th EUROMICRO Conference on Software Engineering and Advanced Applications, SEAA 2011 | Year: 2011

Grid computing enables access to high performance computing resources by offering unified interfaces to compute facilities of distributed compute centers. These compute centers are limited in personal capacity and in support capabilities for different communities/VOs. This work presents a solution to this management problem by facilitating the natural interface introduced by virtualization to separate hardware and application provisioning. Any eScience community may provide preconfigured applications within virtual machine images and compute centers run these without having to maintain applications. This is similar to the Infrastructure as a Service and Software as a Service known from cloud computing. Additionally, in contrast to the state of the art in grid computing, we propose a scalable process to maintain compute resources for scientific users, including the definition of a chain of trust between users and compute centers, which simplifies the addition of communities and resources. As a result, this paper lays the foundation for a next-generation eScience Cloud improving the accessibility of compute resources for research and industry. © 2011 IEEE.

Keller M.,Paderborn Center for Parallel Computing | Kovacs J.,Hungarian Academy of Sciences | Brinkmann A.,Paderborn Center for Parallel Computing
UNICORE Summit 2011, Proceedings | Year: 2011

Research communities from high energy physics to humanities utilised grid infrastructures to support and accelerate their research. Their computations can be executed by different grid technologies: Grids of cluster systems, like the German D-Grid, grids of supercomputers, like the Distributed European Infrastructure for Supercomputing Applications (DEISA), or desktop grids consolidated in the International Desktop Grid Federation (IDGF). UNICORE is one of the three grid middleware environments supported by the European Middleware Initiative (EMI) for managing a set of cluster or supercomputers, but desktop grids are currently unsupported. This work fills this gap enabling UNICORE to support all three kinds of grid technologies of the European Grid Infrastructure (EGI). This unified interface enables European scientists to access web services, portals, and applications on all grid technologies in the same way.

Birkenheuer G.,Paderborn Center for Parallel Computing | Brinkmann A.,Paderborn Center for Parallel Computing | Kaiser J.,Paderborn Center for Parallel Computing | Keller A.,Paderborn Center for Parallel Computing | And 7 more authors.
Software - Practice and Experience | Year: 2012

System virtualization has become the enabling technology to manage the increasing number of different applications inside data centers. The abstraction from the underlying hardware and the provision of multiple virtual machines (VM) on a single physical server have led to a consolidation and more efficient usage of physical servers. The abstraction from the hardware also eases the provision of applications on different data centers, as applied in several cloud computing environments. In this case, the application need not adapt to the environment of the cloud computing provider, but can travel around with its own VM image, including its own operating system and libraries. System virtualization and cloud computing could also be very attractive in the context of high-performance computing (HPC). Today, HPC centers have to cope with both, the management of the infrastructure and also the applications. Virtualization technology would enable these centers to focus on the infrastructure, while the users, collaborating inside their virtual organizations (VOs), would be able to provide the software. Nevertheless, there seems to be a contradiction between HPC and cloud computing, as there are very few successful approaches to virtualize HPC centers. This work discusses the underlying reasons, including the management and performance, and presents solutions to overcome the contradiction, including a set of new libraries. The viability of the presented approach is shown based on evaluating a selected parallel, scientific application in a virtualized HPC environment. Copyright © 2011 John Wiley & Sons, Ltd. Copyright © 2011 John Wiley & Sons, Ltd.

Loading Paderborn Center for Parallel Computing collaborators
Loading Paderborn Center for Parallel Computing collaborators