FinisTerrae

Finisterrae

Location	CESGA Santiago de Compostela (La Coruña) Spain
Coordinates	42°52′33″N 8°33′12″W / 42.875833°N 8.553333°W / 42.875833; -8.553333
Staff	16
Website	cesga.es

Finisterrae was the 100th supercomputer in Top500 ranking in November 2007. Running at 12.97 Teraflops, it would rank at position 258 on the list as of June 2008. It is also the third most powerful supercomputer in Spain (after MareNostrum and Magerit). It is located in Galicia.

This project is promoted by the Xunta de Galicia (regional government of Galicia) and the Spanish National Research Council (CSIC). It was founded in 1993 to serve as a platform to foster scientific innovation and invest in Research and Development.

It is estimated that the base project will be completed in 2010. It is expected to reach the TOP10 of the most powerful supercomputers in the world when it reaches full capacity. The supercomputer is physically hosted at CESGA.

FinisTerrae Distribution Diagram

Overview

The main Finisterrae characteristics are depicted on the following table:

Name	Finisterrae
Place	CESGA
Country	Spain
Year	2007
OS	Linux
Architecture	Itanium 2 64 (Intel)
Processors	2400
Memory	19584 GB
Power	12970 GFLOPS
Top500 Ranking	100
Web	www.cesga.es

One of the special characteristic about this supercomputer is the ratio between cores and RAM. This was one reason why it received the denomination of "singular technical and scientific installation" from the Spanish government, a denomination given to some installations which have some value that makes them singular in some way. Some of those installations include the Canary Island grand telescope, or the Alba synchrotron.

Even if this is the third fastest supercomputer in Spain, some projects that require special amounts of memory cannot be held by the first or second supercomputer, and therefore must be executed on the Finisterrae.

Architecture

FinisTerrae Architecture

HP Integrity Superdome Nodes

HP-SFS storage system Nodes

HP Integrity RX7640 Nodes

HP ESL 712e Tape Library

FinisTerrae supercomputer, located in CESGA is an integrated system by shared-memory nodes with and SMP NUMA architecture. FinisTerrae is composed of

144 computational nodes:

142 hp (106 kW) HP Integrity rx7640 nodes with 16 Itanium Montvale cores and 128 GB memory each one.
1 hp (0.75 kW) Integrity Superdome node, with 128 Itanium Montvale cores and 1024 GB memory.
1 hp (0.75 kW) Integrity Superdome node, with 128 Itanium 2 cores and 384 GB memory.

An hierarchic storing system with:

22 nodes for storing management with 96 processing cores.
The storage is connected to 72 cabins connected to 864 SATA disks, managed by Lustre. Backups are stored on tapes which provide 2.200.000 GB.
An interconnecting network between Infiniband 4x DDR at 20 Gbit/s nodes, providing 16 Gbit/s of effective bandwidth.
An external management Gigabit Ethernet network.
The nodes user SUSE Linux enterprise server as the Operating System. This operating system helped saving an important amount of money when comparing to other non-free options.
FinisTerrae includes open software and standard like: Linux, Lustre, Grid Engine and Globus.
System also has next compilators, libraries and development tools: Intel C/C and Fortran, Intel MKL, Vtune, HP-MPI and HP UPC.

Projects

As main purpose, the aim of the supercomputer is research. The supercomputer is mainly used by the three universities located in Galicia, (Universidade da Coruña, Universidade de Santiago de Compostela and Universidade de Vigo), as well as other research organizations like CSIC. The main projects hold by the supercomputer are divided into four fields:

Life sciences.
Nanotechnology.
Sea sciences.
Renewable energies.

Curiosities

The machines need 140 square meters of space.
It weights around 33,5 tons.
The amount of cables used to interconnect the nodes is around 85 km, the distance between the supercomputer location (Santiago) and Finisterre.

External links, references

Parallel computing

General	Distributed computing Cloud computing High-performance computing

Levels	Bit Instruction Task Data Memory Loop Pipeline

Multithreading	Temporal Simultaneous Preemptive Cooperative

Theory	PRAM model Analysis of parallel algorithms Amdahl's law Gustafson's law Cost efficiency Karp–Flatt metric Slowdown Speedup

Elements	Process Thread Fiber Instruction window

Coordination	Multiprocessing Memory coherency Cache coherency Cache invalidation Barrier Synchronization Application checkpointing

Programming	Stream Processing Dataflow programming Models Implicit parallelism Explicit parallelism Concurrency Non-blocking algorithm

Hardware	Flynn's taxonomy SISD SIMD MISD MIMD Dataflow architecture Pipelined processor Superscalar processor Vector processor Multiprocessor symmetric asymmetric Memory shared distributed distributed shared UMA NUMA COMA Massively parallel computer Computer cluster Grid computer

APIs	Ateji PX Boost.Thread Charm++ Cilk Coarray Fortran CUDA Dryad C++ AMP Global Arrays MPI OpenMP OpenCL OpenHMPP OpenACC TPL PLINQ PVM POSIX Threads RaftLib UPC TBB

Problems	Embarrassingly parallel Software lockout Scalability Race condition Deadlock Livelock Starvation Deterministic algorithm Parallel slowdown

Category: parallel computing Media related to parallel computing at Wikimedia Commons

This article is issued from Wikipedia - version of the 10/21/2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.