July 2011 - aktuell - lists.lrz.de

Upd:News on the HLRB-II successor SuperMUC
by LRZ aktuell July 29, 2011

July 29, 2011

Changes to this message: Corrections to Altix retirement target Dear users of HLRB II at LRZ, The initial installation of the next-generation IBM system with 205 nodes, each containing 40 Westmere-EX cores and 256 GBytes of shared memory, will soon become ready for user operation. This enables us to provide more details about the time table for migration of users from the Altix 4700 to the new system: * We intend to start by selecting large-scale users (job sizes 1000 cores and more) early in August, and incrementally give access also to smaller-scale users throughout August, until all users will be able to access the system in early September * The target is that all HLRB-II users should be able to run their codes on the new system by end of September; in case of difficulties please contact the LRZ Service Desk. * IBM's LoadLeveler will be deployed as a batch queuing system; therefore you will need to make changes to your existing PBS job scripts. * During the intermediate operation phase (from August 2011 to summer 2012), the I/O bandwidth to the scratch disk and project subsystem will be lower than presently available on the Altix, and programs making use of parallel I/O via MPI-IO or HDF5 will require special configuration. * Migration of the LRZ software stack will start in July 2011, and may take until September to be completed. The development environment will be based on Intel's compilers (version 12.0), and either IBM MPI or Intel MPI can be used. Pure OpenMP programs can only be run efficiently with up to 40 threads. * The Altix 4700 will be retired from user operation in early October 2011. * Since there exists no access to scratch and project data produced ($OPT_TMP or $PROJECT) on HLRB-II from the new system, LRZ urgently asks you to delete unneeded data from the scratch and project areas and to write all important data to tape on HLRB-II within the next few weeks. A procedure for restoring TSM tape archives written on HLRB-II to the new system will be provided. Documentation on the new system will be accessible via http:// www.lrz.de/services/compute/supermuc/. Note that as of July 27 2011, this is still work in progress. Further down the road, the following steps will be taken: * In early December 2011, LRZ - in collaboration with Intel and IBM - is planning a training workshop on handling the system and its development environment, to which all users of the HLRB II will be invited. * In summer 2012, the much bigger system (SuperMUC) with more than 6250 nodes (each with 16 Sandy Bridge-EP cores and 32 GBytes of memory) will become available for user operation. SuperMUC will have a peak performance of 3 PetaFlop/s (3,000 TFlop/s) which is roughly 50 times more than the present system. To fully exploit the capabilities of the new SIMD (AVX) units on the Sandy Bridge CPUs, further tuning measures may be appropriate, but programs compiled for the Westmere-based migration system should run without problems. * The system will also provide GPFS ("General Parallel File System") based disk storage with a total capacity of 10,000 TBytes and an aggregate bandwidth of 200 GBytes/s, which will fully support MPI-IO and HDF5-based parallel I/O. Furthermore, SuperMUC will have provisions in place to run jobs in an energy efficient manner e.g., by running processors at a lower frequency if the program's performance does not deteriorate. * Some months after SuperMUC's start of user operation, the nodes of the migration system will be integrated with the main system, forming a "fat node" island for programs which require a large shared memory. We expect that, due to security considerations, some policies on system usage (especially how to access the system) may change. We will keep you informed about these changes through the documentation of the new system once details have been determined. We wish all of you the best possible success for your scientific simulation endeavor on our new Petaflop class successor to HLRB II. If you have further questions, please forward them to us via the LRZ Servicedesk. This information is also available on our web server http://www.lrz-muenchen.de/services/compute/hlrb/aktuell/ali4069/ Reinhold Bader

1 0

Linux Cluster: Ultraviolet presently unavailable
by LRZ aktuell July 28, 2011

July 28, 2011

The Ultraviolet system has run into trouble due to an out-of-memory condition and will require an interruption of service. We're draining the machine and the maintenance is presently scheduled for Friday, July 28. We will keep you informed about the status of the machine via this ALI. Apologies for the disruption of your work. Diese Information finden Sie im WWW unter http://www.lrz-muenchen.de/services/compute/aktuell/ali4071/ Reinhold Bader

1 0

Gauss Call No. 6 for Large Scale Projects
by LRZ aktuell July 28, 2011

July 28, 2011

Details of the Call Large-scale projects and highly scalable parallel applications are characterised by large computing time requirements, not only for short time frames, but frequently for longer time periods. The GCS centres will guarantee that large-scale projects can fully utilise the assigned resources, if necessary at the expense of normal or smaller projects. Projects are classified as "Large-Scale", if they require more than 5% of the potentially available CPU cycles on a member centre's high-end system, i.e. * more than 70 Mill. core-hours at JSC (corresponding to 24 rack months IBM BG/P) * more than 30 Mill. core-hours at HLRS (HERMIT Step 1) * more than 4 Mill. core-hours at LRZ (IBM iDataPlex, Migration System for SuperMUC) For these large-scale projects a competitive review and resource allocation process is established by the GCS. Requests for resources below these limits will be accepted by the individ-u-al member centres. Requests above these limits will be handed over to GCS, handled according to the joint procedures, and will be reviewed in the national context. Projects of smaller scale will be still handled by the local GCS member site. A "Call for Large-Scale Projects" is published by the Gauss Centre twice a year. Dates for closure of calls are usually at the end of February and at the end of August of each year. Eligible are applications from publicly funded academic and research institutions, e.g., universities, Max-Planck Society and Helmholtz Association. The deadline for the current call is 31 August 2011, 17:00. The next call will be issued in January 2012. Answering the Call Leading, ground-breaking projects should deal with complex, demanding, innovative simula-tions that would not be possible without the GCS infrastructure, and which can benefit from the exceptional resources provided by GCS. Application for a large-scale project should be performed by filling in the electronic application form that can be accessed from the GCS web page "Computing Time" (http://www.gauss-centre.eu/computing-time) Note that the regular application forms of the GCS member centres can be reached from there, and users are encouraged to choose a centre that provides the architecture and software most appropriate for their programs. The proposals for large-scale projects will be first reviewed with respect to their technical feasibility. Afterwards, they will be peer-reviewed for a comparative scientific evaluation. On the basis of this evaluation by a GCS committee the projects will be approved for a period of one year and given their allocations. Criteria for decision Applications for compute resources are evaluated only according to their scientific excellence. * The proposed scientific tasks must be scientifically challenging, and their treatment must be of substantial interest. * Clear scientific goals and verifiable milestones on the way to reach these goals. * The implementation must be technically feasible on the available computing system, and must be in proportion to the performance characteristics of these systems. * The Principal Investigator must have a proven scientific record, and she/he must be able to successfully accomplish the proposed tasks. In particular, applicants must possess the necessary specialized know-how for the effective use of high-end computing systems. This has to be proven in the application for compute resources, e.g. by presenting work done on smaller computing systems, scaling studies etc. * The specific features of the high-end computers should be optimally exploited by the program implemen-tations. This will be checked regularly during the course of the project. Further Information * For Questions use: LRZ Helpdesk -Selfservice: https://servicedesk.lrz.de/selfservice/Default.aspx?lang=en&service =Hochleistungsrechnen%20und%20Grid-Supercomputing%20%28HLRB%29 * Current call as PDF http://www.gauss-centre.eu/datapool/page/71/GaussCall6.pdf This information is also available on our web server http://www.lrz-muenchen.de/services/compute/hlrb/aktuell/ali4070/ Matthias Brehm

1 0

News on the HLRB-II successor SuperMUC
by LRZ aktuell July 27, 2011

July 27, 2011

Dear users of HLRB II at LRZ, The initial installation of the next-generation IBM system with 205 nodes, each containing 40 Westmere-EX cores and 160 GBytes of shared memory, will soon become ready for user operation. This enables us to provide more details about the time table for migration of users from the Altix 4700 to the new system: * We intend to start by selecting large-scale users (job sizes 1000 cores and more) early in August, and incrementally give access also to smaller-scale users throughout August, until all users will be able to access the system in early September * The target is that all HLRB-II users should be able to run their codes on the new system by end of September; in case of difficulties please contact the LRZ Service Desk. * IBM's LoadLeveler will be deployed as a batch queuing system; therefore you will need to make changes to your existing PBS job scripts. * During the intermediate operation phase (from August 2011 to summer 2012), the I/O bandwidth to the scratch disk and project subsystem will be lower than presently available on the Altix, and programs making use of parallel I/O via MPI-IO or HDF5 will require special configuration. * Migration of the LRZ software stack will start in July 2011, and may take until September to be completed. The development environment will be based on Intel's compilers (version 12.0), and either IBM MPI or Intel MPI can be used. Pure OpenMP programs can only be run efficiently with up to 40 threads. * The Altix 4700 will be retired from user operation during October 2011, depending on how fast the migration can be concluded * Since there exists no access to scratch and project data produced ($OPT_TMP or $PROJECT) on HLRB-II from the new system, LRZ urgently ask you to delete unneeded data from the scratch and project area and to write all important data to tape on HLRB-II within the next few weeks. A procedure for restoring TSM tape archives written on HLRB-II to the new system will be provided. Documentation on the new system will be accessible via http:// www.lrz.de/services/compute/supermuc/. Note that as of July 27 2011, this is still work in progress. Further down the road, the following steps will be taken: * In early December 2011, LRZ - in collaboration with Intel and IBM - is planning a training workshop on handling the system and its development environment, to which all users of the HLRB II will be invited. * In summer 2012, the much bigger system (SuperMUC) with more than 6250 nodes (each with 16 Sandy Bridge-EP cores and 32 GBytes of memory) will become available for user operation. SuperMUC will have a peak performance of 3 PetaFlop/s (3,000 TFlop/s) which is roughly 50 times more than the present system. To fully exploit the capabilities of the new SIMD (AVX) units on the Sandy Bridge CPUs, further tuning measures may be appropriate, but programs compiled for the Westmere-based migration system should run without problems. * The system will also provide GPFS ("General Parallel File System") based disk storage with a total capacity of 10,000 TBytes and an aggregate bandwidth of 200 GBytes/s, which will fully support MPI-IO and HDF5-based parallel I/O. Furthermore, SuperMUC will have provisions in place to run jobs in an energy efficient manner e.g., by running processors at a lower frequency if the program's performance does not deteriorate. * Some months after SuperMUC's start of user operation, the nodes of the migration system will be integrated with the main system, forming a "fat node" island for programs which require a large shared memory. We expect that, due to security considerations, some policies on system usage (especially how to access the system) may change. We will keep you informed about these changes through the documentation of the new system once details have been determined. We wish all of you the best possible success for your scientific simulation endeavor on our new Petaflop class successor to HLRB II. If you have further questions, please forward them to us via the LRZ Servicedesk. This information is also available on our web server http://www.lrz-muenchen.de/services/compute/hlrb/aktuell/ali4069/ Reinhold Bader

1 0

Skype und NAT-o-MAT/Secomat
by LRZ aktuell July 27, 2011

July 27, 2011

In letzter Zeit kommt es bei der Verwendung von Skype auf Rechnern innerhalb des Muenchner Wissenschaftsnetzes (MWN) immer wieder zu Sperrungen am NAT-o-MAT/Secomat. Das Problem tritt nach unseren Erfahrungen dann auf, wenn der eigene Rechner zum Skype-Supernode und/oder -Relay wird. In diesem Fall steigt die Anzahl der vom eigenen Rechner ausgehenden UDP-Verbindungen rapide an, was vom NAT-o-MAT/Secomat als Port-Scan gewertet und nach einiger Zeit mit einer Sperrung bestraft wird. Dieses Verhalten aendert sich anscheinend von Skype-Version zu Skype- Version, so dass sich keine belastbaren Aussagen ueber einen Zusammenhang zwischen Version und Verhalten machen lassen. Wir haben einen Katalog von Massnahmen in Abhaengigkeit von den jeweiligen Betriebsystemen zusammengestellt, die das Problem schon erfolgreich beseitigt haben: http://www.lrz.de/services/netzdienste/nat-o-mat/#skype_bremse Diese Information finden Sie im WWW unter http://www.lrz-muenchen.de/services/netz/aktuell/ali4068/ Claus Wimmer

1 0

Upd:Linux-Cluster: ICE jobs failing due to network problems
by LRZ aktuell July 27, 2011

July 27, 2011

Changes to this message: Update We are presently observing trouble with the internal network of the ICE system; this causes jobs to fail with strange error messages, segmentation errors etc. The cause of this is presently under investigation. It may become necessary to have an emergency maintenance. We'll keep you up-to-date about the status via this document. This information is also available on our web server http://www.lrz-muenchen.de/services/compute/aktuell/ali4066/ Reinhold Bader

1 0

Netzwartung am 26.7.
by LRZ aktuell July 25, 2011

July 25, 2011

LMU Gebaeude 0030 (Hauptgebaeude mit Turmgebaeude, Bibliothek) Ab 7:30 Uhr werden hier die Switches ausgetauscht. Die Arbeiten koennnen bis ca. 11:00 Uhr dauern. Diese Information finden Sie im WWW unter http://www.lrz-muenchen.de/services/netz/aktuell/ali4067/ Wolfgang Beyer

1 0

Upd:Linux-Cluster: ICE jobs failing due to network problems
by LRZ aktuell July 24, 2011

July 24, 2011

Changes to this message: Update We are presently observing trouble with the internal network of the ICE system; this causes jobs to fail with strange error messages, segmentation errors etc. The cause of this is presently under investigation. It may become necessary to have an emergency maintenance. We'll keep you up-to-date about the status via this document. This information is also available on our web server http://www.lrz-muenchen.de/services/compute/aktuell/ali4066/ Reinhold Bader

1 0

Linux-Cluster: ICE jobs failing due to network problems
by LRZ aktuell July 20, 2011

July 20, 2011

We are presently observing trouble with the internal network of the ICE system; this causes jobs to fail with strange error messages, segmentation errors etc. The cause of this is presently under investigation. It may become necessary to have an emergency maintenance. We'll keep you up-to-date about the status via this document. This information is also available on our web server http://www.lrz-muenchen.de/services/compute/aktuell/ali4066/ Reinhold Bader

1 0

Upd:Einwahl ueber Modem/ISDN funktioniert wieder
by LRZ aktuell July 20, 2011

July 20, 2011

Aenderungen an der Mitteilung: Update Update: Seit Gestern (19.07.) gegen 18 Uhr ist unsere Verbindung zu Vodafone wieder aktiv. Grund des Ausfalls war eine fehlerhafte Verschaltung im Hauptverteiler (Vodafone) Die Modemeinwahl ist nun wieder moeglich. Diese Information finden Sie im WWW unter http://www.lrz-muenchen.de/services/netz/aktuell/ali4059/ Wolfgang Beyer

1 0