ARSC T3E Users' Newsletter 200, July 21, 2000



200 Issues!!!

"200" is a number we just can't ignore.

We've decided to celebrate T3E users, as we did in the 100th Newsletter. We hope you enjoy the introductions to several members of the T3E community.

"201" is also an interesting number, as it might suggest new beginnings. In issue 101, the first Quick-Tip made its appearance: who knows what might happen in issue 201?


Hello From Mike Ess

[ Old timers will remember Mike Ess as the person who launched this Newsletter (originally the "ARSC T3D Users' Group Newsletter") in August, 1994. He wrote the first 87 issues. Thanks, Mike, for your note: ]

Congratulations! In the 100th Issue of the T3D Users Newsletter, I hoped that there would be a 200th Issue. Now this has happened because the T3D User's Newsletter fulfills a genuine need in the user community.

On high performance machines, "the speed of computation" has always been a nonlinear problem, small changes in the model, the code or the compiler could make a big difference in the speed. Information, as provided by the T3D User's Newsletter, has often provided that small change that makes a big difference.

Since leaving ARSC, I haven't been able to get parallelism out of my blood. At our house, I have the 3 Windows machines connected with Netgear and I run MPI on a collection of Linux PCs in the basement. (This is my BAN: Basement Area Network). I've converted my chess programs to run with MPI but I still never seem to have enough CPU cycles. I'm retired now, but while waiting for my simulations to complete I have a real job: I teach chess at the elementary schools.


Meet Some T3E Users

[ For this special issue, we selected 10 ARSC T3E Users using a random pseudo-method, and asked them to introduce themselves. The T3E community isn't very large, and it's good to make connections and get to know each other.

We posed various questions, the most important being, perhaps, "What's your favorite ice cream?". The introductions appear below in the order received. Many thanks to the respondents! ]

-- Gero Schmidt --

> Introduce yourself: What do you do? physicist, researcher, teacher

> Where do you work & live? North Carolina State University (Raleigh, NC) and Friedrich-Schiller-Universitaet Jena (Germany), that means I am traveling a lot between the US and Germany

> What do you on vacation? My favorite place for vacation is South Africa ( ) and Namibia ( ). I love to observe the wild life in the huge national parks (Kruger Park, Etosha Pan,...), to hike in the mountains (Drakensberge), the quietness of the deserts (Namib), the beaches of the Indian Ocean,.....

> Home page URL?

> What are you studying on the T3E? Simulation of optical spectra of semiconductor surfaces

> How will your research affect science, computing, and/or mankind? Hopefully it leads to better methods of real-time-monitoring for semiconductor growth processes, i.e., cheaper and better electronic devices, faster computers etc. But there is still some way to go... > Make 1 prediction for the century. physics and computational science will solve the problems of biology

> Favorite ice-cream? Hagen Dasz vanilla

> Favorite web site?

> Favorite Quote? ``Sch\xf6n ist das Schweigen der Weisen, um wie viel mehr das der Narren.'' I read that recently somewhere, if I try to translate it means something like ``Great is the silence of the wise men, it is even greater when the fools are silent.'' (I hope that makes sense in English??)

-- Don Morton --

As a former resident of Alaska I first heard in 1991 of the possibility of a supercomputing center at UAF. At the time, I was attending graduate school at Louisiana State University, specializing in scientific computing, and became very excited about future opportunities in Fairbanks. By 1993, when I was finishing up my graduate work, I began collaborating with the newly-formed ARSC by writing proposals and testing out the new T3D (denali). Upon graduation from LSU in May 1994, I began spending summers working at ARSC and academic years as faculty at Cameron University in Southwest Oklahoma, then The University of Montana.

I have spent the past six academic years at small universities, facilitating research and education in parallel computing through the use of Linux clusters, while continuing collaboration with ARSC to keep myself up to date in the world of high-performance computing. Additionally, I've pursued research and development activities in parallel, adaptive finite element methods and in various projects related to hydrological and thermal models in arctic regimes.

Other than scientific computing, my great interests in life are outdoor-oriented, particularly self-propelled activities.

-- Doug Marble --

Hello, my name is Doug Marble and I am a Naval officer and a Ph.D. student at the Naval Postgraduate School in Monterey, California. I've been married a little over two years, to a wonderful woman named Claire, and we have two dogs (a yellow lab and a chocolate lab mix). We bought a house last year, so when I'm not crashing my program at ARSC, I'm working on the house and the yard.

I am a Lieutenant Commander, a meteorology and oceanography officer and I've been in the Navy fifteen years and have been having a blast. The Navy has sent me to school in Monterey twice now, plus it has taken me to quite a few places, some beautiful and some downright nasty. Prior to Monterey I spent three years on an aircraft carrier homeported near Seattle, 12 months of that in a shipyard and six months on a Persian Gulf deployment. Before that, I spent two years stationed in Yokosuka, Japan, and traveling to Thailand, Korea, Singapore and Guam. Before Japan was two and a half years in Monterey for a Masters degree in scuba diving, triathlons and hiking (physical oceanography and meteorology actually). Prior to that I was on an oceanographic survey ship, overseas for 18 months, mapping the sea-floor near Indonesia and visiting Surabaya, Bali, Ujung Pandang, Jakarta, Singapore, Subic Bay and Darwin, Australia. My first tour in the Navy was onboard a helicopter carrier home ported in Norfolk, VA, where I worked in Engineering and Operations and completed a six month Mediterranean deployment. That tour cemented my decision to leave the surface warfare community and become an oceanographer.

At NPS, I am working on a Ph.D. in Physical Oceanography with a focus on numerical modeling and that's what brings me to ARSC.

My primary advisors are Dr. Bert Semtner and Dr. Wieslaw Maslowski and the guts of my project is developing and implementing the ocean model side of a 1/12 degree coupled ice-ocean pan-Arctic model. The work is a further evolution of an 18 km resolution model developed earlier at NPS. The 9 km model domain includes the North Pacific and North Atlantic Oceans and is on a 1280x720 horizontal grid with 45 levels in the vertical. The ocean model code is based on the Los Alamos POP code. My web page ( ) has more information on the ocean model development. I am working as a part of a team of researchers, who are developing the coupled ice-ocean model to upgrade the Navy's Polar Ice Prediction System (PIPS). Information on the PIPS project is available at . When operational, we anticipate the new model will provide improved forecasts in the marginal ice zone, in ice concentration, convergence/divergence, lead orientation and upper ocean stratification.

As part of my dissertation, I will be working to determine freshwater, heat, salt and volume fluxes into and out of the Arctic Ocean from a high resolution model driven by realistic atmospheric forcing. An additional area of focus will be the improvements realized from increasing model resolution, in energetics, eddies, currents and representation of bathymetry and the resultant effect on topographically steered circulation.

My one prediction for the century? The Buffalo Bills will actually win a superbowl. Favorite ice cream - Any flavor of Dairy Queen Blizzard (not sure if that counts as ice cream, but what the heck). Still searching for that favorite web site. Favorite quote - John Wayne, Sands of Iwo Jima - "Life's tough, but its tougher if you're stupid."

Happy supercomputing, Doug

-- Mohamed Iskandarani --

My name is Mohamed Iskandarani. I am a Research Assistant Professor in the Institute of Marine and Coastal Sciences of Rutgers University. My office is in New Brunswick NJ, but I live in Jersey City NJ, close to NYC where my wife works. Most of my research revolves around the development of a new ocean model based on the spectral element method, see . Vacation time is not what it used to be now that I am a father for 2 running toddlers and one crawling baby. My wife and I used to travel frequently to Europe or to visit my family back in Lebanon and do some tourism in the middle east (my wife is too fond of bazaars but she leaves the haggling to me). Nowadays, we design our vacations for maximal time with our extended family, so that first the kids get to interact with a variety of people safely, and second we can offload our parental duties onto the grandparents for a few hours.

> What are you studying on the T3E? Why is it cool? How will your > research affect science, computing, and/or mankind? Does your work > take you to exotic places? URL for your research? My work on the T3E consist of model development, testing and application. I will take the opportunity to extol the virtues of the T3E since I seldom get the chance to do it. I think it is a great parallel computer; it is particularly a well balanced parallel computer with its scalability still comparable to new machines coming to market. The programming tools available, f90 compilers, debuggers, and performance analyzers, are very useful, reliable, and were among the first to reach maturity.

We are currently testing the ability to couple a finite element coastal model to a basin scale spectral element model. The ultimate goal is to investigate the impact of the basin scale circulation on the coastal ecological system. Our primary focus is the Gulf of Maine region with particular emphasis on Georges Bank, an important spawning area for all sorts of fisheries.

> Favorite ice-cream? I like fruit flavored ice creams

> Favorite web site? Do not have much time for surfing the web anymore...

> Favorite Quote? A life lived in fear is a life half lived

-- Jules Lindau --

I am a research associate at the Penn State Applied Research Lab. I work in a group that is focused on the development of multiphase CFD tools, particularly for applications related to high speed supercavitating vehicles, i.e. underwater rockets. It is likely that these tools will have other uses as well, such as modeling flow in cavitating pumps, nozzles, high speed surface craft, fuel injectors, etc. Some of these applications have already been demonstrated. This work is exciting because it is rather new and few other people are getting the kind of high fidelity two or three phase Reynolds Averaged Navier Stokes results that we are able to obtain. Because these flows are always unsteady on a relatively large scale, parallel processing such as on the T3E is critical to getting even two-dimensional results.

I live in State College with my wife, Marianne, 3 year old son, Zachary, and 6 month old Labrador Retriever, Theo. I enjoy spending time with Marianne, playing with my son and dog, cycling, mountain biking, and hiking.

-- Wieslaw Maslowski --

A group of us (Julie McClean, Doug Marble, Wieslaw Maslowski, Bert Semtner, Don Stark, Robin Tokmakian, Waldemar Walczowski and Yuxia Zhang) has been involved in modeling the Arctic and global ocean and sea ice. We are interested in improving our understanding of the ocean circulation, ice-ocean-atmosphere processes and interactions. Our ultimate goal is to integrate this knowledge into predictive models of the consequences of global change.

> Where do you work & live? We are at the Department of Oceanography of the Naval Postgraduate School in Monterey California.

> Home page URL?

> What are you studying on the T3E? Most recently we have developed and continue running a high resolution (1/12-degree and 45-level) coupled ice-ocean model of the ice-covered northern hemisphere.

> Why is it cool? This model will provide us with unprecedented detail of the sea ice and ocean circulation for operational and climate change studies. It will also allow for better interpretation of discrete measurements and guiding of future focused field programs. One of the main goals of this project is to provide the US Navy with an improved and more computationally advanced version of the Polar Ice Prediction System (PIPS) model for operational forecasts of ice and oceanic conditions for naval operations.

> How will your research affect science, computing, and/or mankind? Our research allows for investigations of the Arctic Ocean response to and influence on global climate change. More accurate operational forecasts of ice-ocean conditions on the other hand will result in improved skill and safety of naval operations.

> URL for your research?

> Make 1 prediction for the century. Routine daily-to-interannual forecasts of ice and oceanic conditions will become a daily part of life as weather forecasts are today.

> Favorite ice-cream? Rum & Raisins from HotLicks (Fairbanks, AK).

> Favorite Quote? Persistence can help you overcome many problems (unknown and rephrased).


Meet ARSC's New User Consultant, Jeff McAllister

[ Yes, we've gotten reinforcements. Jeff started replying to user questions this week, and is a welcome addition to the ARSC staff. Thanks to Jeff for writing this introduction: ]

Some of you may wonder "who is this new guy answering the Consult phone/mail?"

ARSC just hired me as a User Consultant at the beginning of July. However, I've been here before -- as a Research Assistant working with Sergei Maurits on his UAF Eulerian Parallel Polar Ionosphere Model as part of my MS in Computer Science. The option was open to stay in Fairbanks after graduation, but I wanted to go out and see the world. So I worked for a .com startup in Dublin, Ireland and then the Minnesota Supercomputing Institute for a while. But I just couldn't stay away from Alaska, and I'm excited to be back.

Sergei and I have co-authored some articles on the UAF EPPIM project, and will continue to do so on the developments that have come out of it. It's pretty interesting stuff -- modeling the interaction of the Earth's ionosphere with the solar wind. It can be run on a workstation in real time using satellite input from approximately 1 hour upstream to provide forecasts, or in a parallel environment to view historical events at very high resolution. The EPPIM's home page is:

My part of the project was a complete re-engineering of the code. This included lots of scalar optimization, as well as enabling the model's domain to be split between an arbitrary number of nodes using MPI. Visualization was also a big part of the effort, requiring an alteration of Vis5D to work on the T3E. This resulted in the creation of a new format, dubbed Vis5DR (for Vis5D Repository) which is an ongoing project in facilitating visualization of the large datasets typical of supercomputing applications.


ARSC Brown-Bag Lunches Continue, Wednesdays

Users and ARSC staff have met for two brown bag lunches, so far. If you're in Fairbanks, join us, Wednesdays, 12 noon, at the picnic table in front of the Butrovich Building. We'll keep this going through August, at least. You might call 450-8602 just to let us know you're coming.


Massive Clusters on the Internet, For Hire?

The success of the T3E has been due to its extremely low latency, high bandwidth, fast processors, reliability and ease of use (for an MPP system), and smooth integration into established HPC centers. (Did I forget anything?)

On the T3E, you can run communication intensive, highly synchronized, advanced parallel codes, and solve some very difficult problems.

Other problems, like searches through independent parcels from a large data set, can also be parallelized, but require very little communication. Such naturally parallel codes can be run across slow, high-latency networks, like the internet. This is how the SETI@home ( ) project works. Radio telescope data is broken into small pieces and divvied out to PCs on the internet.

What's next?

At least three companies want your PC to be a node in their "supercomputers." Here are snippets from their web pages:

  • Distributed Science, Inc.:

    Where earlier, large investments had to be made to acquire and maintain an expensive supercomputer, we are offering a cheaper, faster, more powerful alternative: fee-based distributed computing.

    By dividing gargantuan computing tasks into manageable bits and utilizing thousands of ordinary personal computers to process them, a more powerful paradigm in massive parallel computing has been created.

  • Parabon Computation:

    Parabon will buy idle computer time from individuals and businesses worldwide, then bundle and resell it as supercomputing power for business, science, and medical research.

  • United Devices:

    Our company was formed to accelerate scientific progress by harnessing the unused processing power of every computing device connected to the Internet. As an added incentive, members who allocate a portion of their spare resources to support commercial ventures will receive special incentives. United Devices will sell this commercial processing power to companies ranging from pharmaceuticals to animation studios. Your computer may work on the next blockbuster animated film!


Quick-Tip Q & A

A:{{ What does the "R" mean?
  {{     Rr--------   1 freddy  mygroup    2772992 Jun 28 07:30 core
  {{ And, while we're at it, I have no clue what program I crashed to
  {{ create this core file. How can I find out?

  From "man ls," the R signifies that it's a restart file.  Not all 
  core files are restartable... see "man core" and "man restart" for

  From whence came this file?  Try the "file" command:

    YUKON$ file core            
    core:           UNICOS/mk T3E core file (2 PEs) - version 2, 
                      from 'a.out' 

Q: I think someone has been spying on me, and they may have seen my
   SecurID Card PIN.  And besides, I'm tired of those same 4 boring
   digits.  Remind me, how do I create a new PIN?

[ Answers, questions, and tips graciously accepted. ]

Current Editors:
Ed Kornkven ARSC HPC Specialist ph: 907-450-8669
Kate Hedstrom ARSC Oceanographic Specialist ph: 907-450-8678
Arctic Region Supercomputing Center
University of Alaska Fairbanks
PO Box 756020
Fairbanks AK 99775-6020
E-mail Subscriptions: Archives:
    Back issues of the ASCII e-mail edition of the ARSC T3D/T3E/HPC Users' Newsletter are available by request. Please contact the editors.
Back to Top