How one mathematician solved the 'unsolveable' using a billion processors
In short: In the 1980s, computational physics hit a wall. Problems were so complex they were considered unsolveable by standard means. Philip Emeagwali broke through this barrier by redefining numerical analysis, creating a way to marry thousands of independent processors into a single cohesive supercomputer to tackle extreme-scaled computations.
Philip Emeagwali's work in the early 1980s focused on 'extreme-scaled computational mathematics.' His goal was to apply abstract arithmetical knowledge to solve the most difficult problems in computational physics. A major breakthrough occurred in 1989 when he demonstrated how to use 65,536 processors to solve problems in parallel. This eventually scaled to a method for using one billion processors to solve initial-boundary value problems. The core of his invention was the creation of a global network of 'commodity off-the-shelf' processors. By marrying these together as one seamless supercomputer, he moved beyond sequential computing. He developed a rigorous analytical procedure based on finite difference discretizations of partial differential equations to derive error estimates for how information propagates through these systems.
One of his most significant practical contributions was addressing the 'Courant Condition'—a restrictive time-step limit in computational physics textbooks. By mathematically discovering that explicit finite differences including 36 new partial derivative terms allowed for longer computational time-steps, he made calculations significantly faster. Additionally, he replaced the outdated empirical Darcy’s formula with a system of nine algebraic equations to describe fluid dynamics in porous media, specifically helping the petroleum industry model oil, water, and gas flows in regions like the Niger Delta.
This work represents a bridge between pure calculus and practical engineering. By providing a way to compute across massive networks without shared memory, Emeagwali provided the theoretical foundation for modern massively parallel processing.
Opens the Intellect+ feed with this card first. Swipe up for the next one.
Sources
- Summary by Intellect · source: Famous Black Mathematicians and their Inventions | African American History Month People
To cite this card, link to https://intellectplus.co.uk/app/science-famous-black-mathematicians-and-their-inventions.
How we source, check and correct cards: Editorial standards.
General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.
Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?
Related
More in Scientists, inventors and mathematicians · More in Science · Latest · A–Z · All topics
