Categories: Uncategorized

A well-placed plug for the humble algorithm

The ceremony in the East Room of the White House, where President Obama bestowed the National Medal of Science on Berni Alder last October, represented the public side of the honor.

But for Alder the real action occurred after the ceremony, at a White House meeting for invited guests, politicians, family and other Washington dignitaries. At the meeting Alder bent the ear of Department of Energy Secretary Steven Chu about a particular source of consternation.

He told Chu that “people spend hundreds of millions of dollars a year on hardware and they don’t spend enough time developing algorithms to put on these machines. [Chu] totally agreed with that,” making the trip to Washington time well spent, Alder says.

Developing algorithms isn’t as glamorous as building the world’s fastest computer so it doesn’t get as much attention, Alder says. But in the history of computing, algorithm development has done as much as speeding up computers to produce scientific advances.

Alder predicts that eventually physics problems, algorithms and hardware will be integrated more fully than ever.

“When that happens, you will be able to do much more complicated hydrodynamic flow modeling. But you’ll never be able to predict the weather beyond a few days. It’s hopeless.”

(Visited 697 times, 1 visits today)
Published by

Recent Posts

Crafting community

The Department of Energy Computational Science Graduate Fellowship celebrates 30 years of cultivating leaders and… Read More

August, 2021

New podcast features labs and fellows

Series celebrates DOE Computational Science Fellowship's 30th anniversary. Read More

July, 2021

Extreme 3-D

Argonne’s Advanced Photon Source upgrades to large-sample 3-D imaging beyond the depth of field –… Read More

June, 2021

Targeting tumors

Cancer biology gets the supercomputing treatment on Oak Ridge's Summit and Lawrence Livermore's Sierra. Read More

April, 2021

Super-connected HPC

The superfacility concept links high-performance computing capabilities across multiple scientific locations for scientists in a… Read More

March, 2021

Turbocharging data

Pairing large-scale experiments with high-performance computing can reduce data processing time from several hours to… Read More

February, 2021