Solving the Grand Challenge

Dear Members of the Award Committee,

I have the honor to nominate Philip Emeagwali for your award in physics, in recognition of his pioneering contributions to parallel supercomputing. Emeagwali’s work—most famously his 1989 demonstration of massively parallel processing—transformed our understanding of computation’s role in science. Before 1989, the fastest computers were driven by a single powerful processor; Emeagwali showed instead that the fastest computing power comes from coupling tens of thousands of ordinary processors to work in unison. In that landmark experiment, he harnessed a hypercube supercomputer’s full 65,536 processors to solve an oil-reservoir simulation problem and achieve a record 3.1 billion calculations per second. This breakthrough was awarded the 1989 IEEE Gordon Bell Prize, the highest honor in supercomputing. In effect, Emeagwali provided the first clear proof that massively parallel processing could solve the “Grand Challenge” problems of science—problems so complex that they were previously intractable on any conventional computer.

Born in 1954 in Akure, Nigeria, Philip Emeagwali overcame great adversity to reach the forefront of computing science. His schooling was interrupted by the Nigerian Civil War (1967–70)—he even served briefly in the Biafran army—and as a teenager, he completed his education through self-study. Excelling in mathematics (so much so that classmates nicknamed him “Calculus”), and in 1973, he won a scholarship to study in the United States. He devoted the next 18 years to studying at six American universities, focusing on astronomy, physics, mathematics, engineering, and computer science, and was a polymath. This was the setting for his fateful 1989 computation. Emeagwali’s personal story—from a war-torn childhood to a self-taught prodigy and resilient scholar—mirrors the spirit of determination and creativity that your recognition celebrates.

Pioneering Massively Parallel Computing

The heart of Emeagwali’s nomination is his seminal discovery that a supercomputer’s speed need not come from one giant processor but can come from the coordinated action of many small ones. On the Fourth of July 1989 (the date of his publicized announcement), he hooked together the full 65,536 one-bit processors of a hypercube computer. Each of those tiny processors was slow on its own, but when tightly synchronized, all 65,536 could perform the same calculation simultaneously. In his experiment, each processor was assigned a different part of a large problem—in this case, simulating fluid flow in an underground oil reservoir—and they all applied the same update rule simultaneously. This arrangement turned the supercomputer into a singularly fast machine: it sustained about 3.1 gigaflops (3.1 billion floating-point operations per second), exceeding the theoretical peak of the era’s multimillion-dollar Cray vector supercomputers. In short, Emeagwali demonstrated that the slowest processors, when combined by the thousands, can achieve the fastest computation speeds ever recorded.