• Overview• Schedule• Last-Minute Schedule Updates and Changes• My Itinerary• Keynote• Broader Engagement• Cluster Challenge• Education• Important Dates• SCinet• Student Volunteers• SC Fellowship

Click here to check out a comic book from one of the Cluster Challenge teams!

Cluster Challenge

The SC07 committee would like to thank all the teams participating in the Cluster Challenge for their efforts and sportsmanship. Each team and individual on each team should be proud of their achievements last week. Everyone helped make a difference with their activities associated in the Cluster Challenge and hopefully this will be a long-lived event at the SC conferences as a result.

The participating teams and vendor partners included:

Stony Brook University + Dell
National Tsing Hua University (Taiwan) + ASUSTek
University of Colorado + Aspen Systems
University of Alberta (Canada) + SGI
Indiana University + Apple
Purdue University + HP

The winning team was the University of Alberta from Canada with their SGI vendor-partner.

We look forward to the opportunity to run the event again next year at SC08 November 15-21 at the Austin Convention Center in Austin Texas.


Did you know that a small cluster today (less than 1/2 rack) would top the website from just ten years ago? The computational power that is easily within reach today significantly surpasses that available only to the national labs from that time.

The SC07 Cluster Challenge showcases the significance of this and highlights how accessible clusters are to anyone today. In this Challenge, teams of undergraduate students will assemble a small cluster on the Exhibit floor and run benchmarks & applications selected by industry and HPC veterans. Taking place Nov. 10-16, 2007, at the Reno-Sparks Convention Center in Reno, NV, the Cluster Challenge will be judged on the speed of benchmarks and the throughput of application runs over the first three days of the conference.



The Cluster Challenge is a showcase event in which teams of next-generation high performance computing talent harness the incredible power of current-generation cluster hardware. This challenge brings together an international field of teams that share a “need for speed” and a willingness to earn the top prize. The event promises to be exciting, educational and a truly rewarding experience for all involved.

In the spring of 2007, contest application software used for the challenge will be announced. Teams may study and tune open source benchmarks and applications for their platforms (within the rules, of course).

During SC07 in Reno, teams will assemble, test and tune their machines until the green flag drops on Monday night as the Exhibit Opening Gala is winding down. The race now begins and teams are given data sets for the contest. With CPUs roaring, teams will be off to analyze and optimize the workload to achieve maximum points over the next two days.

In full view of conference attendees, teams will execute the prescribed workload while showing progress and science visualization output on large displays in their areas. As they race to the finish, the team with the most points will earn the checkered flag – presented at the awards ceremony on Thursday.

After the checkered flag drops, teams are invited to partake in the side-show, where they can spin their wheels and show off what they’ve learned and what they can do with the equipment. There are no rules, just curiosity as visitors will be looking for activities that defy gravity, solve the Ramsey number problem or perhaps even reverse global warming.



A team is six (6) individual student members, a supervisor and optional vendor partners. The students provide the brain-power, the vendors provide the systems and other (non-technical) support and the supervisor brings…. pizza!

To qualify for a team, an individual must not have been granted an undergraduate degree (as of the start of the contest). Younger team members, perhaps in high school, are welcome, but must meet the minimum age requirements of the SC07 show. Among a number of safety rules, contestants will be limited to a maximum of 12 hours per day in the contest area.

The required supervisor must be an employee of the team’s educational institution. The supervisor will need to be present and available 24x7 for the duration of the show. While the supervisor is not allowed to provide technical assistance, he/she is encouraged to run for fuel (Hot Pockets, pizza and soda) for their team.

Teams are expected to partner with vendors who may support team activities by providing equipment and funds for travel. The event coordinators will provide a list of unmatched teams and vendors to assist in contact making.

Special local housing arrangements and other logistical support is being arranged for teams. Limited funds may be made available to assist in travel.


Clusters are to be provided by the team and must consist of a single full-height 19” rack. A monitoring power strip will be available into which all components of the cluster must be plugged. A single 30 amp, 110 volt circuit will be provided with a soft cap at 26 amps. Alarms will be sent electronically if power draw exceeds this amount and penalties may be assessed for excess draw.

The chosen hardware must be commercially available and teams must display (for public view) a complete list of hardware, software and any provided services and the associated costs. Prices shown must be honored should random visitors wish to make a purchase.

In addition to the competition hardware, teams must also provide a 42” display and showcase their progress on this display throughout the week. Laptops are allowed to locally access the cluster, monitor activities, and drive the display.

A network drop will be provided for outgoing connections only. Offsite access to the equipment will not be permitted.


Teams may choose any OS and software stack that will run the benchmarks and application software.

For the main portion of the challenge, teams must run the HPC Challenge benchmarks (found at These results must be published prior to starting on the scientific applications.

Scientific applications will be chosen to be OS-neutral and to provide real-world workloads for clusters of the size anticipated for this event. The applications will be announced prior to the submission deadline, giving teams time to ponder what they are getting themselves into. Teams may study the applications and modify them for their platforms in advance of the event.

When the green flag drops, teams will be handed media with data sets for the applications. While teams execute the HPC challenge benchmarks, they may want to study the data sets and plot a strategy for the next three days. Points are awarded for successful processing of data sets and displaying output on the monitors for visitors to follow.

Good luck!

The benchmarks and applications for the Cluster Challenge have been chosen to be portable across a variety of expected operating systems and hardware platforms. The benchmarks are widely in use as a metric for measuring system performance, and the applications are also widely used in their specific disciplines.

It is not necessary to fully understand the science behind each application to succeed in the Cluster Challenge. The applications will, in general, compile and run on any system, and require only standard mathematical and message-passing libraries. What may help your team, at the time of the challenge, is an ability to analyze the input and understand the computational resources required for a given data set.

When the Cluster Challenge begins, you must first run the benchmarks (see below for details). Once you have completed this task, and registered your results with the judges, you may move on to the applications. Once you have started on the applications, you may not return to the benchmarks.

Note: Keep an eye on this page for updates and additional information

Benchmarks: High Performance Computing Challenge – HPCC Benchmark

For the past 21 years, since 1986, the high performance computing community has relied on a single benchmark, LINPACK, to rate their systems. LINPACK solves a dense system of linear equations. The historical and current listings of high-performance systems are available on the top500 website at Recently the HPCC benchmarks (, which include LINPACK, have been growing in popularity. The Cluster Challenge will use this benchmark suite.

HPCC was developed to study future Petascale computing systems, and is intended to provide a more realistic measurement of modern computing workloads. HPCC is made up of seven common computational kernels: STREAM, HPL, DGEMM (matrix multiply), PTRANS (parallel matrix transpose) , FFT, RandomAccess, and b_eff (bandwidth/latency tests). The benchmarks attempt to measure high and low spatial and temporal locality space. The tests are scalable and can be run on a wide range of platforms, from single processors to the largest available parallel supercomputers.

The HPCC benchmarks test three particular regimes: local or single processor, embarrassingly parallel, and global, where all processors compute and exchange data with each other. STREAM measures the memory bandwidth of a processor. HPL is the LINPACK TPP (Toward Peak Performance) benchmark. RandomAccess measures the rate of random updates of memory, PTRANS measures the rate of transfer of very large arrays of data from memory, and b_eff measures the latency and bandwidth of increasingly complex communication patterns. All of the benchmarks are run in two modes: base and optimized. The base run allows no source modifications of any of the benchmarks, but allows generally available optimized libraries to be used. The optimized benchmark allows significant changes to the source code. The optimizations can include alternative programming languages and libraries that are specifically targeted for the platform being tests.

The reference implementation of the benchmark is available at A “C” compiler and an implementation of MPI are required to run the benchmark. The report titled Introduction to the HPC Challenge Benchmark Suite by Dongarra and Luszczek ( describes how HPCC was used at SC06.

The Parallel Ocean Program (POP)

Overview (from

POP is an ocean circulation model derived from earlier models of Bryan, Cox, Semtner and Chervin, in which depth is used as the vertical coordinate. The model solves the three-dimensional primitive equations for fluid motions on the sphere under hydrostatic and Boussinesq approximations. Spatial derivatives are computed using finite-difference discretizations which are formulated to handle any generalized orthogonal grid on a sphere, including dipole and tripole grids which shift the North Pole singularity into land masses to avoid time step constraints due to grid convergence. Time integration evolves the ocean model and can predict the evolution of properties of the oceans such as sea-surface height variability.

POP is freely available to the community (under a copyright agreement); you can download the latest version from these pages. For more information and to download the latest version please see


Overview (from

The General Atomic and Molecular Electronic Structure System (GAMESS) is a general ab initio quantum chemistry package. It can be used to compute the properties of molecules and chemical reactions using a wide variety of theoretical models. These include equilibrium geometries, vibrational frequencies, with IR or Raman intensities. The Fragment Molecular Orbital method permits use of many of these sophisticated treatments to be used on very large systems, by dividing the computation into small fragments. The code can run serially or in parallel and graphics programs. The MacMolPlt program (for Macintosh, Windows, or Linux desktops, see, is available for viewing the final results, and the Ghemical program can assist with preparation of inputs. The input is a simple pseudo-formatted file and the main output is a text file as well. Graphics programs automatically parse the output to produce images.

Special Download Information: When you go to download GAMESS you will have to agree to the user agreement and then get additional information about how to download GAMESS via email. There is no cost for GAMESS but you have to agree to the terms of the user agreement. The download pages describe an optional program TINKER. For the SC07 Cluster Challenge, TINKER will not be required. You will want the source code distribution of GAMESS even though many binary packages are available. The main GAMESS page is

POV-Ray The Persistence of Vision Ray-Tracer (

The Persistence of Vision Ray-Tracer creates 3D, photorealistic renderings using a technique called “ray tracing”. The images are created from simple text instructions that describe the geometry of the objects in the scene, the lighting, surface characteristics and other relevant data such as the viewpoint and special effects. The extremely high quality of the images generated comes at the cost of computation.

Ray tracing is a general method, which originated in the field of geometrical optics. The concept, as applied to 3D computer animations, traces, in reverse, a path that could have been taken by a ray of light, which would intersect with a viewpoint. As the scene is traversed by following the rays in the inverse direction, visual information (color) about the scene is accumulated from the point of view or eye point. As the ray intersects objects in the scene, additional rays can be generated to represent reflection, refraction, or adsorption/shadowing.

POV-Ray is freely available as either binary or source. Versions for Unix, Linux, Windows and MacOS X can be downloaded from Its use is subject to an end-user license. The code requires a C compiler and does not use any explicit parallel execution environments. Animations are typically built up from scenes rendered on individual CPUs of a cluster.

     Home  |  About  |  Contact Us  |  Registration ACM    IEEE    The Computer Society