LRZ-Newsletter Nr. 05/2012 vom 03.05.2012
Unsere Themen:
-
Zentrale Systeme und Grid Computing
- PRACE Call for Proposals
- PRACE Preparatory Access Call for Tier-0
- PRACE Summer School on Code Optimisation for Multi-Core and Intel® MIC Architecture
- Gauss-Allianz-Infobrief: GA-Infobrief
- PARA 2012: Workshop on State-of-the-Art in Scientific and Parallel Computing
- ISPDC 2012 - Eleventh International Symposium on Parallel and Distributed Computing in Garching/Munich, 25-29 June 2012
-
Stellenangebote / Job Opportunities
Impressum, Abbestellen des Newsletters und Archiv
Termine, Veranstaltungen, Allgemeines
Leibniz-Rechenzentrum erhält "Deutschen Rechenzentrumspreis 2012"
Der "Deutsche Rechenzentrumspreis 2012" in der Kategorie "Energie- und Ressourceneffiziente Rechenzentren" ging am Abend des 29. März 2012 bei der Gala "future thinking" in Sinsheim an das LRZ.
In seiner Bewerbung "Höchstleistungsrechnen – höchst effizient!" hatte das LRZ die weltweit einzigartige Energieeffizienz seines neuen Höchstleistungsrechners "SuperMUC", der zur Zeit installiert wird, herausgestellt. Durch die innovative Kühlung des Systems mit warmem Wasser ist es möglich, den Großteil der Rechnerkomponenten ganzjährig ohne den Einsatz von Kältemaschinen zu kühlen. Die effiziente Ausnutzung der Energiesparmechanismen neuester Prozessortechnologien erlaubt zudem die gleichzeitige Optimierung der Rechnerperformance für Geschwindigkeit und Energie. Auch auf diesem Gebiet ist das LRZ Pionier.
Weitere Einzelheiten und ein Foto finden Sie auf unserer Webseite zur Verleihung des "Deutschen Rechenzentrumspreises 2012"
WLAN-Verbesserung im Maschinenwesen der TUM
Durch die stark zunehmende Zahl von Studierenden sowie der mobilen Geräte war die WLAN-Versorgung im Gebäude des Maschinenwesen in Garching nicht mehr ausreichend. Die Studienbeitragskommission der Fakultät Maschinenwesen hat deshalb knapp 70.000 Euro für die Verbesserung bereitgestellt. Gemeinsam mit dem LRZ wurde ein Konzept erarbeitet und teilweise bereits umgesetzt.
In einem ersten Schritt wurden in allen öffentlichen Bereichen die bestehenden APs (Accesspoints) durch neueste Modelle mit einer Bandbreite von bis zu 400 Mbit/s ersetzt. In den Hörsälen, in denen die Anzahl gleichzeitiger Nutzer naturgemäß noch höher ist, wurden zusätzliche APs zur Verstärkung installiert. Insgesamt wurden über 70 APs verbaut.
Durch die im letzten Jahr extrem zunehmende Zahl an Nutzern wurden auch die für WLAN verwendeten IP-Netze zu klein. Da die bisherigen Switches in den Gebäudehauptverteilern nur 30 verschiedene Netze (VLANs) unterstützten und die Lehrstühle, Rechnerpools, etc. jeweils auch eigene Netze benötigen reichte dies nicht mehr aus. Als Konsequenz mussten in den acht Gebäudehauptverteilern die Zentralswitches erneuert werden.
Da aus Brandschutzgründen keine zusätzlichen APs in der Magistrale installiert werden dürfen, konnte die Planung des LRZ in diesem Bereich nicht umgesetzt werden und die Versorgung in der Magistrale ist deshalb noch nicht optimal. Es ist jedoch eine Lösung erarbeitet worden, welche nur passive Antennen im Fluchtwegbereich vorsieht. Wegen aufwändiger Leitungsinstallationen ist die Umsetzung aber voraussichtlich erst bis zum Beginn des Wintersemesters 2012/2013 möglich.
Der Zeitplan für die erledigten Arbeiten war sehr ambitioniert. Die Mittel wurden Mitte Februar bewilligt, und die Ersetzung konnte komplett in der vorlesungsfreien Zeit bis zum 15. April abgewickelt werden. Dies war auch nur möglich, weil im LRZ neben den verantwortlichen Mitarbeitern drei Auszubildende zusätzlich mit dem Löwenanteil ihrer Arbeitszeit für das Projekt gearbeitet haben.
Autor: Wolfgang Beyer
Zentrale Systeme und Grid Computing
PRACE Call for Proposals for 5th Project Access and Multi-year Access (Tier-0) and synchronized Call for DECI-9 (Tier-1)
PRACE (Partnership for Advanced Computing in Europe) is a Research Infrastructure that allows researchers to apply for high-performance computers time from a series of hosting nations via a central peer review process. This call is the fifth PRACE-RI Regular call for Project access, inviting applications for high-end (Tier-0) computing resources to carry out projects which have high scientific quality and impact. Allocation will be for 1 year starting from 1st November 2012.
This Call is also open to applications for Multi-year Projects lasting up to 2 years. Allocation of resources is made one year at a time, with provisional allocations awarded for the second year. All Multi-year Access Projects are subject to annual peer review, based on progress report, and the allocation of resources shall be adjusted accordingly.
The call also invites proposals for project access to Tier-1 resources, via DECI, the Distributed European Computing Initiative, providing single project cross-national access to national (Tier-1) HPC-resources. The deadline for submission of proposals both for Tier-0 and for Tier-1 is 30th May 2012 at 12:00 CEST.
Opening date: 17th April 2012
Closing date: 30th May 2012, 12:00 CEST Start date: 1st November 2012 Allocation period: 1 year from 1 Nov 2012 to 31 Oct 2013 Type of access: Project access and Multi-year access
For more information please see: http://www.prace-ri.eu/
PRACE Preparatory Access Call for Tier-0
Preparatory access is intended for testing and developing codes in order to prepare applications for PRACE Tier-0 project access. In particular, we would like to draw your attention to the Preparatory Access Type C which includes code development with support from PRACE experts. Further information is given on the PRACE web page http://www.prace-ri.eu/Call-Announcements .
The Preparatory Access Call is a rolling call, researchers can apply for resources all year. There are no closing dates, proposals will be reviewed every third month. The next cutoff date is June 1, 2012.
LRZ invites all interested scientists, whose projects fit the criteria of the PRACE calls to submit their projects. Further information about PRACE calls as well as information on the PRACE machines, can be found at http://www.prace-ri.eu/ .
If you need some extra information please feel free to contact PRACE directly via
peer-review@prace-ri.eu or
weinberg@lrz.de .
This information is also available on our web server http://www.lrz-muenchen.de/services/compute/aktuell/ali4297/ .
Autor: Volker Weinberg
PRACE Summer School on Code Optimisation for Multi-Core and Intel® MIC Architecture
The Partnership for Advanced Computing in Europe (PRACE) is pleased to announce the Summer School on Code Optimisation for Multi-Core and Intel® Many Integrated Core (MIC) architecture at the Swiss National Supercomputing Centre in Lugano, Switzerland. In this three day intensive event participants will focus on programming and tuning techniques for modern multi- and many-core processors with a particular focus on the Intel® Many Integrated Core architecture. Topics will be presented at an advanced level and will include structuring code to enable SIMD/vectorization, efficient usage of register, cache and memory hierarchies, use of multi-threading techniques to maximize resource utilization, data locality considerations on multi-socket NUMA nodes, and inter-node communication. Intel specialists will introduce the Intel® MIC architecture and the Intel programming environment, and will delve into greater detail of the use of multi- and many-core programming techniques. Demonstrations and hands-on sessions will be integrated throughout the course to illustrate the topics in greater depth, and user case studies will highlight real-world experiences in optimising scientific (HPC) codes on the Intel® MIC architecture.
Dates
21st-23rd June 2012
Location
Swiss National Supercomputing Centre, Lugano, Switzerland
Registration
To apply for a place please go to http://survey.ipb.ac.rs/index.php?sid=43493&lang=en. The school is offered free of charge to students and academics residing in PRACE member states, and lunches, coffee breaks and an evening dinner event will be provided. Please note that a sound knowledge of C, C++ or Fortran is a pre-requisite for this course. Attendance is limited to 30 places, and selection of participants will be based on demonstrated understanding of computational science and high-performance computing and experience with at least one of MPI, OpenMP, or multi-core programming. Registration closes Friday 11th May 2012, but we encourage applicants to apply as early as possible. Candidates will be informed of the results of their application by the second week of May (at the latest).
Registration deadline: May 11,2012.
Please contact tim.robinson(at)cscs.ch for further information or see http://prace-portal.cscs.ch/blog/?tx_typopress_pi1[post]=138 .
Autor: Alexander Block
Gauss-Allianz-Infobrief: GA-Infobrief Nr. 7
Ein neuer Infobrief der Gauss-Allianz ist erschienen:
- Strategische Weiterentwicklung des HPC in Deutschland
- Föderatives Identitätsmanagement in Baden-Württemberg
- CUDA Center of Excellence in Dresden
- Skalierbare ELPA-Eigenwertlöser als Hybridversion
- Interaktive Visualisierung von HPC-Datensätzen
- Monitoring und Optimierung von E/A-Vorgängen
- EU investiert in energiesparenden Exascale-Rechner
PARA 2012: Workshop on State-of-the-Art in Scientific and Parallel Computing
Helsinki, Finland, June 10-13, 2012:
The main theme of PARA12 is State of the Art in Scientific and Parallel Computing with a special theme "Towards Exascale". Interesting topics include (but are not restricted to) algorithms, software, tools, environments as well as applications for scientific computing, high-performance computing, grid computing, cloud computing, and interactive visualization.Important dates and deadlines
- Registration ends May 31, 2012
- PARA 2012 Tutorials June 10, 2012
- PARA 2012 Workshop June 11-13, 2012
- Deadline for proceedings papers September 1, 2012
For more details please see the CSC webpage.
ISPDC 2012 - Eleventh International Symposium on Parallel and Distributed Computing in Garching/Munich, 25-29 June 2012
The mission of the 11th International Symposium on Parallel and Distributed Computing is to bring together researchers, engineers and practitioners involved in the development of new algorithms and architectures for parallel and distributed computing. The goal is to provide an interactive and friendly yet professional forum for original research contributions describing novel ideas, groundbreaking results and/or quantified system experiences. We cordially invite you to submit your contribution to the 11th International Symposium on Parallel and Distributed Computing (ISPDC 2012).
Stellenangebote / Job Opportunities
Scientific Computing Experts
Leibniz Supercomputing Centre (Leibniz-Rechenzentrum, LRZ) is part of the Bavarian Academy of Sciences and Humanities (Bayerische Akademie der Wissenschaften). LRZ is a central site for large scale data archiving and backup. It is an active player in the area of high performance computing for over 20 years and provides computing power on several different levels. It offers its services to universities in Germany and to publicly funded research institutions, like Max-Planck and Helmholtz Institutes. LRZ is currently preparing the installation of its new high-end system (SuperMUC) with a peak performance of 3 PetaFlop/s and more than 150,000 cores. When SuperMUC goes into operation it will be a world-leading platform for accelerating open scientific discovery.
For the support of the users and applications on the new system LRZ is looking for
Scientific Computing Experts
The positions are available immediately.
Required experience and skills
- PhD or university (post) graduate degree, diploma, master's degree or equivalent in a scientifically oriented subject (computational science, mathematics, physics, chemistry, life sciences etc.)
- Good HPC skills, a broad knowledge of massively parallel high-performance algorithms, programming languages and libraries as well as performance analysis tools.
- Initiative, capacity for teamwork, and creativity.
- Good oral presentation and writing skills in German and English
Preferred Experiences
- Experience with large-scale parallel application development, debugging and performance tuning
- Demonstrated ability to lead scientific and technical efforts in HPC
- Experience in managing scientific projects and/or teams
Duties/Responsibilities
- Perform work in the application support team
- Optimization, scaling and performance improvement of strategic large-scale applications
- Performance analysis and benchmarking
- Identification and solution of user problems
- Keeping contact with scientific stakeholders and users
- Promotion and organization of scientific outreach activities through presentations and providing written materials
- Purchasing and maintaining third party applications and tools
- Writing of reports, grant applications, and documentation
- Stimulation of new developments in HPC
Due to its state-of-the-art computing facilities and its proximity to the Munich universities, LRZ offers a very stimulating and innovative environment. Progressive policies such as flexible working hours support the combination of work and family life.
We offer a pleasant and easygoing work environment in an internationally recognized HPC team with many years of experience that you can draw on for support, while also having ample opportunity to work independently and follow individual preferences.
Salary is determined by the German trade agreement for public service "Tarifvertrag für den öffentlichen Dienst der Länder" (TV-L.). The position is time-limited, initially, to two years with the expectation of prolongation. Qualified women are strongly encouraged to apply. Severely disabled persons will be hired preferentially over others with equal qualifications. Due to export control of supercomputers restrictions for nationality may apply. For further information please contact:
Dr. Matthias Brehm
Group Leader HPC
Phone: +49 89 35831 8773
Email: brehm_at_lrz.de
Submission of your application via email is preferred. Please submit it with resume/C.V., certificates, reports and letters of recommendation to:
Leibniz Supercomputing Centre
Mrs. Hannelore Apel
Keyword: HPC-Experts 2012/01
Boltzmannstr. 1
85748 Garching near Munich
Germany
Email: apel_at_lrz.de
Impressum
Herausgeber:
Leibniz-Rechenzentrum der Bayerischen Akademie der Wissenschaften
Anschrift:
Leibniz-Rechenzentrum der Bayerischen Akademie der Wissenschaften
Boltzmannstraße 1
D-85748 Garching
Telefon: +49-89-35831-8000
Telefax: +49-89-35831-9700
E-Mail: lrzpost_AT_lrz.de
Redaktion: Dr. Ludger Palm
Bestellen und Abbestellen des Newsletters
Sie können den LRZ-Newsletter über unsere Webseite bestellen oder abbestellen.
Archiv
Frühere Ausgaben des LRZ-Newsletters finden Sie im Archiv.