IBM Blue Gene
Supercomputers go green
Overhead view of the tall black boxes of the Blue Gene supercomputer system, with 8 rows of units lined up in a server room

IBM’s computers have consistently ranked among the most powerful in the world. In the early 1960s, the IBM 7030 Stretch reigned as the world’s fastest computer, capable of processing more than 1 million instructions per second. Four decades later, another IBM computer held the crown — the IBM Blue Gene. It could perform trillions of operations per second.

As fast as it was, Blue Gene revolutionized the field of supercomputing for another reason. It ushered in an era of high-performance computing that valued efficiency as much as raw processing power.

Developed in collaboration with the US Department of Energy’s Lawrence Livermore National Laboratory in California, the first-generation Blue Gene, known as Blue Gene/L, was geared to help biologists observe the processes of protein folding and gene development. It had more than 65,000 computing nodes. Each node contained a microchip that held an array of processors, mathematical engines, memory and communication systems. The densely packed, power-efficient machine took up far less space than its ancestors and so opened up supercomputing to many more researchers hoping to gain a deeper understanding of how proteins fold and diseases develop. 

Subsequent versions, Blue Gene/P and Blue Gene/Q, enabled universities, governments and commercial research labs to address a wide range of problems that had simply been too complex to tackle previously. Blue Gene systems have simulated radioactive decay, replicated brain power, flown airplanes, pinpointed tumors and modeled emerging climate trends.

Doing more with less

For half a century, building a powerful computer meant constructing an enormous machine that inevitably required a tremendous amount of energy. The Electronic Numerical Integrator and Computer (ENIAC), introduced in 1945, was the first programmable, electronic, general-purpose digital computer and weighed 30 tons, covered 1,500 square feet of floor space and used more than 17,000 vacuum tubes. Running it required 160 kilowatts of electrical power. Even at that size, ENIAC could execute only about 5,000 additions per second.

In the decades that followed, high-performance computers became exponentially faster but still demanded an enormous amount of energy and square footage — and produced so much heat that elaborate cooling facilities had to be constructed to ensure proper operation. All of which proved bad for the environment and made owning a supercomputer incredibly expensive.

In 1999, a team of IBM engineers and scientists looked to chart a new course for supercomputer design with the Blue Gene/L prototype. From an engineering standpoint, the guiding principle of the team was simple but innovative: Do more with less.

IBM’s Alan Gara was chief system architect for the three generations of Blue Gene supercomputers. He identified power consumption and reliability as two of the primary constraints on the continued scaling of supercomputing architecture. Gara created a design based on low-power system-on-a-chip (SoC) nodes, with dense packaging and multiple interconnection networks that scaled beyond anything previously envisioned. The design dramatically simplified the number of instructions carried out by each processor, enabling them to work faster and with significantly lower power and chip surface requirements. Gara’s groundbreaking work won the IEEE Computer Society’s 2010 Seymour Cray Award.

Blue Gene took five years and USD 100 million to develop, with the first machine developed at IBM’s Thomas J. Watson Research Center in New York and built at the company’s facility in Rochester, Minnesota. The name of the computer was a nod to its work with genetics combined with IBM’s nickname, “Big Blue.” On September 29, 2004, the IBM Blue Gene/L became the fastest computer in the world, surpassing NEC’s Earth Simulator. Blue Gene/L used 131,000 processors to handle 280 trillion operations every second. One scientist with a calculator would have to work nonstop for 177,000 years to perform the operations that Blue Gene could do in a single second.

Blue Gene computers went on to enable significant advances in designing ultra-efficient electric car batteries, understanding global climate change and exploring the evolution of our universe.


On September 29, 2004, The IBM Blue Gene/L became the fastest computer in the world
Green supercomputing

As it has turned out, the color most commonly associated with the Blue Gene line of supercomputers is green. And its design has charted a new course for the computing industry. Today, many of the world’s most energy-efficient supercomputers are built on IBM high-performance computing technology. In 2010, Blue Gene/Q topped the Green500 List, ranking as the “Greenest Supercomputer in the World.” President Barack Obama recognized the Blue Gene family of supercomputers with the National Medal of Technology and Innovation, the country’s most prestigious honor for technological achievement.

For decades, computer performance was judged solely on processing speed, as measured in floating-point operations per second, or flops. Increasingly, supercomputer performance is measured in flops-per-watt, a recognition of the importance of energy efficiency in supercomputer design — an advance pioneered by Blue Gene.

In 2010, Blue Gene/Q ranked as the “Greenest Supercomputer in the World”
Related stories Sustainability

A founding commitment to make the world a better place lives on at IBM today

Breaking the petaflop barrier

IBM Roadrunner smashed an elusive speed record and marked a new era in supercomputing

The IBM 7030, aka Stretch

With ‘Stretch,’ IBM entered the age of supercomputing