The Leading Source for Global News and Information Covering the Ecosystem of High Productivity Computing
February 29, 2008
Simulation Reveals Secrets of the Origin of the Universe
1 | 2 All »
Our understanding of structure formation in the universe has grown at an astonishing rate over the past twenty years. On the one hand, observations carried out by satellites and telescopes have given us a more precise knowledge of the "initial conditions" of our universe, providing us with a sound basis for its dynamic evolution over its 13 to 14 billion year existence. And on the other hand, the dramatic increase in available computing resources now allows increasingly complex algorithms to simulate the formation of the structures of the universe.
As part of the "Horizon Project," a team of French scientists, led by Romain Teyssier, an Astrophysicist at CEA (the French Atomic Energy Commission), has completed the largest simulation ever carried out of structure formation in the universe. This simulation will enable astrophysicists to compare their models with astronomical observations with an unprecedented level of realism. The aim of the project is to get a better understanding of the moment when the universe started to form structures under the influence of gravity: how small primordial density fluctuations grow by gravitational instability and finally form galaxies and galaxy clusters we observe today.
To achieve this ambitious objective and simulate such a volume of information with a high enough level of detail, extremely powerful computing resources and highly efficient algorithms were both needed -- as well as a lot of expertise and know-how to make them work together in an optimum way! The members of the Horizon project therefore ran the RAMSES fluid dynamics code on 6144 Intel Itanium cores of the Bull NovaScale cluster installed at CCRT. Thanks to a close collaboration between the scientific team, the CCRT systems team, and Bull's HPC experts, the program was able to use the resources of the cluster in an optimum way for almost two months, consuming 18 terabytes of memory and generating nearly 50 terabytes of data on disk.
The CCRT cluster currently consists of 932 Bull NovaScale 3045 compute nodes, each equipped with four dual-core Intel Itanium 2 processors -- making a total of 7456 compute cores. It is managed via two administration nodes and 24 I/O (input/output) nodes, is based on a high-performance Voltaire InfiniBand DDR 4X interconnect network, and has a storage capacity in excess of 420 TB. This supercomputer has a peak performance of approximately 50 Tflops, and is one of the five systems in the world offering the best performance ratio (Rmax/Rpeak) in the Top50.
The deployment of the cluster was quite an achievement in itself, since the system went into operation only five months after the order was placed -- just in time to secure a good position in the June 2007 TOP500 ranking. The cluster software stack is resolutely open source-oriented, and is based on a Linux system optimized by Bull for large-scale computers. It includes powerful administration tools, such as the Bull NovaScale Master suite which controls each cluster component -- a particularly useful feature when code must run efficiently for as long as two months!
To maximize system performance, the compute software suite was specifically optimized by Bull for this type of computer, with:
- A kernel optimized for the Itanium 2 /SMP architecture.
- A MPIBull2 library (Message Passing Interface: communication library) delivering 5us latency over the entire cluster.
- A powerful Lustre file system with a W/R throughput of more than 22 GB/s on 10 DDN storage bays. The performance of the file system was a key success factor in achieving the RAMSES simulation.
The RAMSES code was developed at CEA's Astrophysics Department by Romain Teyssier to study large-scale structure and galaxy formation. It is a flexible package that can be used for general-purpose simulations in self-gravitating fluid dynamics. The parallel code is written in Fortran 90 with extensive use of the MPI (Message Passing Interface) library for communications between processors. The Horizon project was an opportunity to enhance the scalability of the software, thanks to a close co-operation with Bull's HPC experts.
RAMSES is a grid-based hydro solver. Each cell in the grid can be "refined," i.e., divided into smaller cells to obtain a finer description of the system at a local level. This refinement can be applied repeatedly to any cell, which enables very high resolutions to be achieved in the specific locations, where necessary. This technology is known as Adaptative Mesh Refinement (AMR).
As far as the Horizon simulation is concerned, this technology enabled the team to achieve unprecedented spatial resolution (equivalent to a cubic grid with a side length of 262,144 grid points!). The universe was divided up into elementary meshes, which were subdivided into smaller meshes in places where the density of matter exceeded a predefined level. With nearly 70 billion black matter particles and over 140 billion meshes, the Horizon simulation is an absolute record for a computer-simulated N-body system (for more information, see: http://www.projet-horizon.fr). It describes half the observable universe, with enough resolution to describe a Milky Way-like galaxy with more than 100 particles.
"With this new simulation, we will be able to predict how matter is distributed in the universe with great precision and realism," explains Romain Teyssier. "We will soon be in a position to compare our model with full-sky observations, such as the one that will be shortly available thanks to the Planck space mission being launched by the European Space Agency in 2008. We will also be able to prepare future experiments on gravitational lensing, such as the 'Dark UNiverse Explorer' (DUNE), a project aimed at determining the nature of dark energy."
Page: 1 of 21 | 2 All »
Article Tools
Share Options
(Digg, Technorati, more)
Subscribe
Sponsored Links
RSS Feeds
Feeds by Topic
- Applications
- Developer Tools
- Interconnects
- Middleware
- Networks
- Processors
- Storage
- Systems
- Visualization
Feeds by Industry
Feeds by Content Type
Feature Articles
Spider Up and Spinning Connections to All Computing Platforms at ORNL
Spider, the world's biggest Lustre-based, centerwide file system, has been fully tested to support Oak Ridge National Laboratory's new petascale Cray XT4/XT5 Jaguar supercomputer and is now offering early access to scientists.
Read More...
Wolfram Alpha: A Web-Based Application That Embraced Supercomputers
Wolfram Alpha, the Web-based computational engine introduced in May, is not a traditional supercomputing application, but relies on supercomputers to satisfy its unique requirements.
Read More...
TeraGrid '09: Student Participation Soars
There was a new energy at this year's TeraGrid '09 conference thanks to an outstanding turnout for the student program. Thanks to support from the National Science Foundation, more than 100 high school, undergraduate and graduate students were able to participate in the conference.
Read More...
Top Headlines
3D Seismic Data: Taking a Smarter Approach to Interpretation
Jul 09 | Engineer Live | The demand for computational tools to underpin the 3D seismic interpretation process has never been more apparent. Read more...
Engineering Unemployment Soared in 2Q to 8.6%
Jul 08 | EE Times | Unemployment for U.S. engineers has reached record levels, according to government figures. Read more...
Gartner Adjusts 2009 IT Spend Downward Again
Jul 08 | Network World | Global spending for 2009 projected to drop 6 percent, for a total of $3.2 trillion. Read more...
Concurrent and Parallel Are Not The Same
Jul 08 | Linux Magazine | Portability or efficiency? Neither is guaranteed when writing explicit parallel code. Read more...
800 TFLOP Real-Time Ray Tracing GPU Unveiled, Not for Gamers
Jul 07 | Ars Technica | Japanese company builds custom ASIC to accelerate real-time ray traced rendering for the auto industry. Read more...
Featured Whitepapers
Building High Performance Computing in a Green and Modular Solution Building Block
Apr 14 | | Many HPC IT departments are feeling the rising pressure to deliver more capacity computing and performance while trying to reduce the total cost of ownership. This white paper discusses how an environmentally-friendly and open-standards HPC building block based computing system using flexible interconnect options helps address capacity computing needs.
Multimedia
Webcast: Dell Expands HPC Access and Adoption with Intel Cluster Ready Program
Source: Addison Snell, GM/VP, Tabor Research; sponsored by Dell
Many organizations that could benefit from the use of HPC clusters find that it is complicated to get the systems up and running because of limited IT resources or the complexities of the clusters themselves. Learn how the Intel Cluster Ready program, for which Dell was an original partner, seeks to address this challenge for entry level and mid-range HPC users.
Video White Paper: Architecting a Better Network Storage Solution
BlueArc's Titan architecture represents an evolutionary step in file servers by creating a hardware-based file system that can scale bandwidth, IOPS, and overall data capacity well beyond conventional software-based devices. With its ability to virtualize a massive storage pool of up to four usable petabytes of tiered storage, Titan can scale with growing data requirements, offering a competitive advantage for businesses, researchers, or other enterprises seeking to better manage data growth while still ensuring optimal performance.
Webcast: HPC Development Solutions: Sun Studio & Sun HPC ClusterTools
Sun Studio Compilers and Tools and Sun HPC ClusterTools allow you to create high performance parallel applications for OpenSolaris, Solaris and Linux. Sun Studio Express 11/08 includes MPI performance analysis capabilities and full OpenMP 3.0 compiler support. Learn about all this and the latest in Sun HPC ClusterTools 8.1.
Newsletters
Stay informed! Subscribe to HPCwire email Newsletters.
HPC Job Bank
-
HPC Systems Administrator
McGill University -
High Performance Computing Data Center Manager
National Renewable Energy Laboratory -
High Performance Computing Systems Administrator
Idaho National Laboratory -
Solution Engineer- HPC
Dell Inc -
Parallel Programmer
Cardiff University -
Principal Software Research Engineer
Indiana University -
Post Doctoral Researcher - Exascale Stream Computing
IBM - Visit the HPCwire Job Bank
Featured Events
-
July 13-16, 2009
WORLDCOMP'09
Las Vegas , NV
USA
-
August 24-28, 2009
Data Mining: Levels I, II, & III
Denver , CO
USA