Beschikbare supercomputers
Supercomputers in Nederland
Voor technisch/wetenschappelijk gebruik staat in Nederland een aantal supercomputers klaar voor gebruik. Via NCF kan voor gekwalificeerd onderzoek toegang tot een of meer van deze systemen verkregen worden. Toegang is primair bedoeld voor onderzoekers aan Nederlandse instellingen voor wetenschappelijk onderzoek, in het bijzonder universiteiten en wetenschappelijke instellingen zoals NWO- en KNAW-instituten.
Overzicht van systemen
De volgende systemen zijn in Nederland beschikbaar voor landelijk gebruik:
- IBM systeem (HUYGENS)
- Nationaal Cluster (LISA)
- IBM BlueGene/P (STELLA)
- GPU Cluster at SARA
Toegang aanvragen
IBM P575 cluster (HUYGENS)
Huygens is an IBM Hydro-Cluster system, consisting of SMP (Symmetric MultiProcessing) nodes. In July 2008 the second phase of Huygens has been installed and made available to users in August 2008. The system consists of 1664 dual-core POWER6 processors, so 3328 POWER6 CPU cores. The total amount of central memory is approximately 15.25 TB, with a net amount of disk storage capacity of around 700 TB, distributed across scratch space and home file systems. The amount of nodes of the Huygens system is 104, each node with a peak performance of 32*18.8 Gflop/s, so 601.6 Gflop/s, which leads to a system peak performance of 104*601.6, which equals 62.6 Tflop/s. 18 nodes have been equipped with 256 GB of main memory, the other nodes have 128 GB of main memory. One node acts as the central node, available for interactive access, compilation and submitting of jobs to the batch system. Two nodes act as test nodes for new software versions, which makes the amount of nodes available for batch processing equal to 101. Simultaneous Multi Threading (SMT) is activated standard, which means that one node of 32 cores can run 64 threads or processes in an efficient way. Each core runs at 4.7 GHz, which makes the peak performance per core at 4*4.7 = 18.8 Gflop/s (4 floating-point operations per cycle). Each node runs the Linux Operating System All nodes have been connected through a low-latency, high-bandwidth InfiniBand interconnect network, with an MPI latency of 4.5 µs and 4.8 Gbyte/s bandwidth between neighbouring nodes.
Contactpersonen bij NCF: Peter Michielse of Ana Bela Sa Dias
Nationaal Rekencluster (LISA)
SARA has installed the National Compute Cluster on behalf of:
- The University of Amsterdam (UvA)
- Free University Amsterdam (VU)
- Netherlands National Computing Facilities Foundation (NCF)
- NWO institutes (through NCF)
- Van ’t Hoff Institute for Molecular Sciences
The system has been configured as follows:
- 536 compute nodes:
- 120, each equipped with:
- 2 Intel® quad-core Xeon" E5345 processors
Speed: 2.33 GHz
Extended Memory 64 Technology (EM64T) - 1330 MHz Front Side Bus
- Gigabit Ethernet adapter for maintenance
- 2 Intel® quad-core Xeon" E5345 processors
- 32, each equipped with:
- 2 Intel® quad-core Xeon" L5420 processors
Speed: 2.5 GHz
Extended Memory 64 Technology (EM64T) - 1330 MHz Front Side Bus
- Gigabit Ethernet adapter for maintenance
- 2 Intel® quad-core Xeon" L5420 processors
- 384, each equipped with:
- 2 Intel® quad-core Xeon" L5520 processors
Speed: 2.26 GHz
8M cache 5.86 GT/s QPI, Turbo - Gigabit Ethernet adapter for maintenance
- 2 Intel® quad-core Xeon" L5520 processors
- Topspin InfiniBand network:
- Bandwidth: 800 MB/sec
- Latency: <6 µsec
- Qlogic InfiniBand network::
- Bandwidth: 1600 MB/sec
- Latency: <6 µsec
- Total peak performance: 19 TFlop/sec
- Disk space: 48 TB SUN® Thor systems for the home file systems
- Operating system: Debian Linux AMD64 OS
There are five types of compute nodes:
Number Processors Memory InfiniBand 40 2*quad-core 8GByte Yes 80 2*quad-core 8GByte No 32 2*quad-core 16GByte Yes 128 2*quad-core 24GByte No 256 2*quad-core 24GByte Yes - 120, each equipped with:
Plaats: SARA, Amsterdam
Contactpersonen bij NCF: Peter Michielse of Ana Bela Sa Dias
IBM BlueGene (STELLA)
The Astron Lofar/CIT Blue Gene P system consists of 3072 compute nodes, each with 2 GB of memory. One compute node consists of 4 cores, providing a total of 12288 cores. A number of “I/O nodes” is used for the communication with the compute nodes and for performing the I/O operations between compute nodes and data servers. Although the system can be used as one massive compute block it has been divided in logical “partitions”. These partitions communicate with the frontend of the system via their I/O nodes and can be used separately. Jobs submitted to disjoint partitions can be run at the same time. The LOFAR data processing operates almost continuously and does not take part in any scheduling or job submission of the scientific part. The part available for the scientific users operates as a batch processing system: Jobs are submitted to a queue are matched with a compute block when a suitable partition becomes available. The scheduling system tries to ensure efficient usage of the machine. Note that the Blue Gene is a real batch processing system which implies that jobs can only start when others have been terminated. On the compute nodes only one process can run per processor at the same time. The CIT DNA queueing system on the system controls job submission and scheduling. The DNA (Dynamic Node Allocation) system controls access to the partitions via a number of “queues”. Queues contain a predefined maximum number of cores and maximum compute time. Users may be allowed to submit jobs to more than one queue. Once a job has been submitted to a queue the DNA system schedules the job and will run it on a partition which can supply the number of requested cpu’s when they become available.
Plaats: CIT-RUG, Groningen
Contactpersonen bij NCF: Peter Michielse of Ana Bela Sa Dias
GPU Cluster at SARA
The GPU cluster is a cooperative project of SARA Computing and Networking Services, the University of Amsterdam and the Netherlands National Computing Facilities Foundation (NCF).
The GPU cluster consists of the following hardware:
Each node contains:
- 2 quad CPU processors: Intel(R) 2.5 GHz
- 32 GByte memory
- 233 GByte scratch disk z
- 2 NVIDIA(R) Tesla graphic adapters, with the following specifications
The GPU cluster is embedded in the Lisa cluster. The CUDA programming environment of NVIDIA is installed. Since the GPU cluster is embedded in the Lisa cluster, the same software as on the Lisa system is available.
Plaats: SARA, Amsterdam
Contactpersonen bij NCF: Peter Michielse of Ana Bela Sa Dias
Voor alle machines is meer gedetailleerde informatie (in het Engels) beschikbaar via de site www.supercomputer.nl , SARA (gebruikersinformatie) of bij RUG-CIT.
Hoe kan toegang worden aangevraagd?
Aanvragen kunnen via Iris worden ingediend (zie ook onder NCF-subsidie-formulieren). Iris is het elektronisch aanvraagsysteem van NWO, dat de interne afhandeling van aanvragen sterk bespoedigt. Door op onderstaande link te klikken komt u in de NWO subsidiewijzer terecht.
Rekentijd aanvragen via Iris.
