Content about Technology

Feature - Dash heralds new form of supercomputing

Dash, pictured here, is an element of the Triton Resource, an integrated data-intensive resource primarily designed to support UC San Diego and UC researchers. Image courtesy of San Diego Supercomputer Center, UC San Diego.

The first of a new breed of supercomputers was born this fall when computer experts combined flash memory with supercomputer architecture to create Dash.
Normally supercomputers are measured by how many floating point operations, also known as “flops,” they can complete per second. And at a peak speed of 5.2 teraflops, Dash wouldn’t even make the top 500 list, where the slowest speed is about 17 teraflops.
“But if you look at other metrics, such as the ability to do input/output operations, it would potentially be one of the fastest machines,” said Allan Snavely, the project leader for Dash. “Dash is going after what we call data-intensive computing, which is quite different from

Q & A - Smart data handling: An interview with Tevfik Kosar

Image courtesy Tevfik Kosar

In e-science, we are constantly striving to improve performance and speed so that we can complete a larger number of more complex computations faster. Tevfik Kosar, a researcher at Louisiana State University, is working on two intertwined projects that could together lead to the sorts of improvements we hope for. Read on to find out what he had to say.
iSGTW: You just received a National Science Foundation grant to work on the Stork Data Scheduler. Can you tell us a little about that project?
Kosar: The funding for the development of Stork Data Scheduler comes from NSF's Strategic Technologies for CyberInfrastructure program. The STCI program funds innovative cyberinfrastructure services which have the potential to significantly advance research capabilities in multiple areas of science and engineering. The grant will provide three years of funding for the enhancement of the Stork Data Scheduler

Advanced registration deadline for ICPADS09

The Shenzhen skyline. Image courtesy of IEEE ICPADS 2009.

Advanced registration for the Fifteenth International Conference on Parallel and Distributed Systems (ICPADS09) closes on 1 October, after which prices will rise by over $100.
This year the event will take place in Shenzhen, China, 8-11 December 2009.
According to the event's website, ICPADS 2009 provides an international forum for scientists, engineers, and users to exchange and share their experiences, new ideas, and latest research results on all aspects of parallel and distributed systems.
Keynote speakers include Kai Hwang from the University of Southern California, Xiaodong Zhang from the Ohio State University, Sajal K. Das from the University of Texas at Arlington, and Denis Caromel from INRIA Sophia Antipolis-CNRS.
 
 
 

Feature - New batch of science gateways hit the spot

A visualization of a gauge configuration generated at NERSC which is now freely available via Gauge Connection, a NERSC science gateway that serves as an experimental gateway for the lattice quantum chromodynamics community. Image courtesy of NERSC.

Accessing high performance computing resources via the web can be as easy as everyday tasks such as paying bills, shopping and chatting with friends, thanks to a new Science Gateways project. But it wasn’t always that way.
The traditional method of accessing computing center resources is to log in, write and submit a mini-computer program called a batch script, and then wait for the results. Science gateways allow researchers to accomplish the same tasks using a web-based graphical user interface (if it can use a mouse, it’s a graphical interface). This means that scientists can dive right into the science without worrying about learning how to write a batch script.
“Th

Link of the Week - Deutsche Welle

Deutsche Welle is Germany’s international broadcaster: online, on-screen and over the air. It provides a European perspective to audiences around the world and promotes intercultural dialogue. Image courtesy Deutsche Welle

The man who founded Sun Microsystems and wrote the first check for a start-up called Google now wants to build 100-gigabit switches for the huge data centers of the looming cloud.
Andreas von Bechtolsheim explained his reasoning, saying: “Historically, people literally had to buy a computer to then use the program or the application they wanted for the business, payroll, database and so on. In the future, all of these applications will run just on the web in some big remote data centers located in the places of the world where the power is cheapest and the user only needs a browser on the laptop or on the mobile phone to access these applications.”
To find out more, check out what he told Deutsche Welle, our Li

Opinion - Multiple challenges for multicore processors

Quad-Core AMD Opteron processor. Image courtesy AMD

Since the microprocessor’s advent over 30 years ago, the vast majority of software applications have been built and executed on single processor computer systems. We have lived through an age of easy programmability where large numbers of software developers have learned to program in a variety of languages which can easily be compiled into machine code for the latest, greatest microprocessor.For many years, high performance computing followed the same path, with faster, more-complex processors delivering annual performance gains following Moore’s Law. By the mid-1980s it became clear that single-processor performance improvements couldn’t continue to be delivered, and parallel computing was born. The transition from serial to parallel programming is difficult, and as a result the number of programmers trained to develop code in this way has remained less than 1%

Link of the week - The Dagdigian Doctrine

Image courtesy of Bio-ITWorld.com

What's old news? What's exciting now? What's down the road? What's just hype?In the opening keynote of the 2009 Bio-IT World Conference & Expo last April in Boston, Chris Dagdigian told it as he saw it. A founding partner for the BioTeam IT consultancy, he delivered an educated assessment of the best, the worthwhile, and the most overhyped information technologies  for the life sciences.Ken Davies covers Dagdigian's talk in Bio-ITWorld.com. You'll only find a few highlights from the article here . . .What does Dagdigian think is dead?“Unlimited data storage is over,” he noted, according to Davies. Dagdigian asserted that it is simply not possible to back up all data and keep it safe. “Sometimes it’s better to go back to the -40 F freezer and repeat the experiment.”What's hot now?  He offered virtualization as the “lowest hanging fruit.” Not coming soon,

Feature - Busting the myth of immutable microprocessors “Many Paths” © Sue O'Kieffe 2008 A whimsical representation of the myriad paths the logic can take through a reconfigurable microprocessor Imagine reconfiguring your computer’s hardware to perform individual tasks as efficiently as possible.  It's new, it's green, and it's . . . well . . . challenging. A standard computer’s hardware structure is fixed. To accomplish different tasks, its microprocessor runs different software programs. In a reconfigurable computing system — an up-and-coming technology pioneered by the National Science Foundation Center for High-performance Reconfigurable Computing (CHREC, pronounced “shreck”) and others — the hardware structure actually changes when performing different tasks. Reconfigurable hardware is proving capable of providing substantial improvements in power reduction and speed, in particular when applied to high-performance comp

Feature - Embrace failure!

Don Lamb addressing the Fault Tolerance for Extreme Scalability Workshop, co-sponsored by The National Science Foundation Office of Cyberinfrastructure’s Blue Waters and TeraGrid projects.
Image courtesy of TeraGrid External Relations.

Can smart checkpoints and fault-resilient applications avert a Malthusian Catastrophe?
As more powerful systems encompass ever-increasing numbers of components, even a small fault rate on individual processors will generate multiple faults across the components, stopping long-running applications in their tracks.At a workshop in March, U.S. experts met to discuss issues relating to the fault-tolerance of today’s and tomorrow’s petascale and exascale computing systems. The group explored past practices and common pitfalls, and discussed strategies to ensure that these systems and the applications they run can tolerate the inevitable faults. “It is invaluable for the systems specialists, middlew

Feature - Let scientists focus on the science David Abramson (at right) and colleagues show how Nimrod can be applied to scientific problems. Image courtesy of Monash University. David Abramson, Professor of Computer Science at Monash University in Australia, and Director of the Monash eScience and Grid Engineering (MeSSaGE) Lab, contributed this article. When scientists find the complexity of dealing with high performance computers too high, they miss out on the advantages it has to offer. An understandable view — their expertise is science, not coercing a computer into doing what they want.  Over the years, our group has helped many scientists and engineers, in disciplines ranging from quantum chemistry to public health policy, to embrace new computing technologies that will advance their research.This started about 15 years ago. We suggested to a group of physicist colleagues that they run their code in parallel on a cluster. Their problem separated in a straightforward manner into independent, parallel tracks, and a

Feature - Grids and Clouds: GridBriefing just released Image courtesy of GridTalk So, we’ve all heard about clouds and grids. But what are the pluses and minuses of each approach? For that matter, just what exactly is a “cloud?” A just-published paper goes into this and more, with comparisons of grid computing and cloud computing. Written by former iSGTW editor Cristy Burne (now of GridTalk) and available as a GridBriefing at the GridTalk website, the four-page document dissects the benefits and drawbacks of each approach, and draws upon opinions and sources as diverse as Hewlett Packard and the European Software Association. There is also a quick-reference chart that goes into such meat-and-potatoes issues as: Who provides the service?Who uses the service?Who pays for it?Where are the computing resources?Why use them?What are they useful for?How do they work? Grids and clouds, side-by-side. Click image to read chart.  Image courtesy GridTalk.   Pointers The GridBriefing has pointers to furthe

Opinion - Many Task Computing: Bridging the performance-throughput gap In the lower left corner, the low number of tasks and small input size makes tightly-coupled Message Passing Interface (MPI) quite manageable. This is the traditional terrain of HPC.  As the data size increases (vertically), we move into the analytics category, such as data mining and analysis.  In the lower right corner, data size remains modest, but the increasing number of tasks moves us into loosely-coupled applications involving many tasks. HTC can be considered a subset of this category. Finally, the combination of both many tasks and large datasets in the upper right corner moves us into the province of Many-Task Computing. MTC can also be considered as part of the high-task, low data (lower right) area. Image courtesy of Ioan Raicu, University of Chicago. Tightly-coupled applications for which jobs must communicate between each other during execution are typically best served by clustered High Performance Computing (HPC). Applications with many