Turing Award Won by Programmer Who Pioneered Supercomputers


In the late 1970s, as a young researcher at the Argonne National Laboratory outside of Chicago, Jack Dongarra helped write the computer code called Linpack.

Linpack offered a way to run complex math on what we now call supercomputers. It has become a vital tool for scientific laboratories as it expands the boundaries of what a computer can do. This included predicting weather models, modeling economies, and simulating nuclear explosions.

On Wednesday, the Computer Machinery Association, the world’s largest community of computing professionals, announced that 71-year-old Dr. He said Dongarra will receive this year’s Turing Award for his work on the core concepts and code that allows computer software to keep up with hardware around the world. The most powerful machines in the world. The Turing Prize, which has been awarded since 1966 and often referred to as the Nobel Computer Prize, comes with a prize of $1 million.

In the early 1990s, Dr. Dongarra and his colleagues have created a new type of test that can measure the power of a supercomputer using the code Linpack (short for linear algebra package). They focused on how many calculations it could make per second. This has become the primary way to compare the fastest machines in the world, grasp what they can do and understand how they should change.

“People in science often say, ‘If you can’t measure it, you don’t know what it is,'” said Paul Messina, who heads the Department of Energy. Exascale Computing Project, a software development effort for the nation’s best supercomputers. “That’s why Jack’s job is important.”

Now a professor at the University of Tennessee and a researcher at nearby Oak Ridge National Laboratory, Dr. Dongarra was a young researcher in Chicago when he majored in linear algebra, a form of mathematics that underlies many of the most ambitious tasks in computer science. . This includes everything from computer simulations of climates and economies to artificial intelligence technology. meant to imitate the human brain. Developed with researchers at several American labs, the software library called Linpack has helped researchers run this math on a wide variety of machines.

“Basically these are algorithms you need when dealing with problems in engineering, physics, natural science or economics,” said Ewa Deelman, a professor of computer science at the University of Southern California who specializes in software used by supercomputers. “They let the scientists do their job.”

Over the years, as we continue to develop and expand Linpack and adapt the library for new types of machines, Dr. Dongarra also developed algorithms that could increase the power and efficiency of supercomputers. As the hardware inside the machines continues to evolveSo is the software.

In the early 1990s, scientists disagreed on the best ways to measure the progress of supercomputers. Thus, Dr. Dongarra and colleagues created the Linpack benchmark and began publishing a list of the 500 most powerful machines in the world.

Updated and published twice a year Top 500 list It sparked a competition between scientific labs to see who could build the fastest machine — bypassing the gap between the “ball” and the “500”. What began as a battle for bragging rights gained an added advantage as laboratories in Japan and China challenged traditional strongholds in the United States.

Dr. “There’s a direct parallel between the computing power you have within a country and the types of problems you can solve,” Deelman said.

The list is also a way of understanding how technology is evolving. In the 2000s, he demonstrated that the most powerful supercomputers were computers that combined thousands of small computers into one huge whole, each equipped with the same types of computer chips used in desktop computers and laptops.

In the years that followed, Amazon followed the rise of “cloud computing” services from Google and Microsoft, connecting smaller machines in even greater numbers.

These cloud services are the future of scientific computing, as Amazon, Google, and other internet giants are building. new kind of computer chips that can train AI systems speed and efficiency that was never possible in the pastDr. Dongarra said in an interview.

“These companies are making chips that are tailored to their needs, and that will have a huge impact,” he said. “We will rely more on cloud computing and ultimately abandon the ‘big iron’ machines in national labs today.”

Scientists also a new kind of machine called a quantum computerThis can make today’s machines look like toys by comparison. As the world’s computers continue to evolve, they will need new benchmarks.

Dr. “Manufacturers will brag about these things,” Dongarra said. “The question is: What is truth?”



Source link

Leave a Reply

Your email address will not be published.