Emeagwali
|
philip Emeagwali is not a normal Nigerian. Even after amassing degrees in five different fields, his name still comes bald with no attachments - not even a mere Dr. Coming from a country where even engineers (Engineer Ololo) and other lesser mortals (Dr Mrs Chief Alhajia Olusoga; Ambassador Baba Kuti) prefix their names with a chaotic mix of titles, Emeagwali's eschewal of the national fixation shows why, in the words of the American magazine Michigan Today, he is "one of the world's fastest humans."
In 1989 he became the first man on earth to perform the world's
fastest computer computation - a staggering 3.1
billion calculations per second. That surpassed even the theoretical
peak speed of the more expensive supercomputers
costing between $10m and $30m a piece then in use in the US.
Emeagwali's achievement also helped solve one of America's 20 Grand Challenges --- understanding how oil flows underground so companies could extract the most of the "black gold." Emeagwali's breakthrough won him the prestigious Gordon Bell Prize in the US. Though worth only $1,000 in cash and awarded by the American Institute for Electrical and Electronics Engineers, the Gordon Bell is considered to be the Nobel Prize of computing.
Emeagwali captured it while making mince-meat of the previous winners' efforts. His 3.1 billion calculations per second was twice the speed achieved by the 1988 winner and 24 times faster than the 1987 winner. And what's more - he was the first solo winner of the award. Usually, teams from corporations and national laboratories capture the $1,000 cheque. But Emeagwali, a "novice" by American computing standards at the time he started his project, spent only eight months on a relatively new type of computer - the Connection Machine - before clinching the award. The Nigerian embassy in Washington milked the applause: "By this great accomplishment," Dr Hamzat Ahmadu the then Nigerian ambassador in Washington wrote to Philip, "you have shown the world that, contrary to many negative press [reports] that tend to portray the generality of Nigerians in bad light particularly in the USA, Nigerians are capable of achieving greatness and do indeed excel in many fields. I hope that other Nigerians both within and outside your community will emulate you and set similar high standards of excellence."This is what Nigerians at home call "reaping where you've not sown." For, Emeagwali achieved his greatness without any help from the Nigerian government. At age 14, Philip dropped out of school in Onitsha where he grew up because his father, James, (a nurse) could not pay his school fees. Philip, an Ibo born at Akure in Yorubaland, was the first of his father's nine children. The Nigerian government, despite all the oil money, did not help the poor boy to continue his education. But before he dropped out of school, Philip had shown a clear head for mathematics. His father James was so encouraged that he continued to teach him mathematics in the evenings. Philip was such a mathematical genuis that his mates soon nicknamed him "Calculus." Even his father, the evening teacher, could not keep up with him. "He gave up because he said I knew more than he did," Philip told New African. Having dropped out of school, Philip made the public library his second home where, through library books, he taught himself college-level mathematics, physics, chemistry and English. He studied hard and passed his GCE from the University of London with ease.
His working life has seen him in Maryland and Wyoming as a civil engineer, and as a computer scientist at the US National Weather Service for which he wrote a thesis on mathematical calculations used in forecasting floods. He also worked at the US Army High Performance Computing Research Centre in Minneapolis from 1991 to 1993. He has since won several awards including the Scientist of the Year Award 1991 given by the National Society of Black Engineers in the US ("in recognition of his outstanding contribution to the scientific field which benefitted all mankind"); the Computer Scientist of the Year Award 1993 given by the National Technical Association; and the Pioneer of the Year Award 1996, again given by the National Society of Black Engineers.
The project was for his doctoral dissertation. Instead of using one supercomputer which cost $30m a pop and uses eight high-powered processors, the Nigerian resorted to the Internet to access 65,536 smaller computers each working in parallel on a small part of the problem. "The result was phenomenal: a computer that performed computations more than three times faster than a supercomputer," wrote Marsha Lakes Matyas for the American Association for the Advancement of Science. Emeagwali was estactic: "I like to work on problems that are important to society because you get satisfaction," he told New African. "Research is hard work, so you might as well work on important research". He tells how he did it. "The Connection Machine had only been developed in 1986, three years before my breakthrough, and it had a poor reputation. People were sceptical about the Machine and denigraded it because, typically, researchers needed a year to learn how to use it and its software was scarce. "I was one of the first persons crazy enough to try it. I'm more confident in pursuing ideas that are not very well accepted. It has something to do with my personality. If everyone goes east, I go west, so I will be the first to find the solution." But not many believed in what he was doing. "To be really honest, I thought he was getting caught between the cracks," confessed Trevor Mudge, Director of the Advanced Computer Architecture Laboratory at the University of Michigan where Emeagwali studied. But the Nigerian knew what he was doing. He linked up 65,536 processors - each comparable to a desk top computer - via the Internet to the Connection Machine. He figured that more points would provide better results. He says: "Just picture the conventional supercomputer, costing $30m each, as eight oxen pulling a cart and the Connection Machine as about 65,000 chickens pulling the same cart. "The old thinking is that the oxen will do a better job, but if the chickens coordinate their efforts, they'll do a better job than the oxen. That's what my 65,536 small computers achieved: 3.1 billion calculations per second. The supercomputer was nowhere in sight with all its $30m purchase price."
In the US, the government has a list of the 20 most difficult
problems in science and engineering called the 20 Grand Challenges.
One of them is petroleum reservior simulation. Emeagwali
wanted to crack this.
He explains: "Oil is usually found underground, trapped in rocks.
Until my experiment, en-gineers were able to recover only about 10%
of the oil in any reservior by using supercomputers to simulate
oil fields and track the oil flow.
I chose this problem because in this field, in order to have the
most impact, you have to work on the most serious problems."
Coming from Nigeria, he knew that an oil field was not just a
huge underground cave.
"Oil is found in pores within rocks," he continues, "and oil
companies must pump gas or water into fields to force the oil
to nearby wells.
"But if the oil is sucked out too quickly at one well, then oil
elsewhere may not flow naturally to the same well and is virtually
trapped. The company must then drill another well at considerable
expense to extract the trapped oil.
"The money involved is staggering. That's why I set myself to
understand the flow within oil fields and how many barrels can be
removed at any one well".
To help him overcome the problem, Emeagwali looked elsewhere.
He knew that in the early 1930s, the German mathematician Paul
Fillunger had formulated a theory on partial differential equations.
But Fillunger was not "successful." A sceptical panel of experts
publicly announced in March 1937 that the German had failed to
solve the equations. Fillunger felt so humiliated that he and his
wife committed suicide. After their deaths, the Russian mathematician
B. K Risenkampf seized on Fillunger's idea, but he too was only
able to partially solve the equations. But Philip Emeagwali showed
in 1989 that Fillunger's original equations were correct.
He modified Fillunger's equations on his computer and then
divided the simulated oil field into eight million points,
assigning 128 points to each of the 65,536 processors linked
to the Connection Machine.
He had written his programme to instruct each point in the
65,536-chain to talk with six neighbours simultaneously.
The results were phenomenal!
if he were
Archimedes, the great Greek mathematician, he would
have run out of his bath shouting "eureka,
eureka" (I've found
it, I've found it).
But Emeagwali was Nigerian. When the Connection Machine popped up
the results, his eyes went wild! He screamed and punched the air.
He had achieved 3.1 billion calculations per second! And not only
that: His calculations had determined the amount of oil in the
simulated oil field, its direction of flow and speed at each point.
For the entire oil field, (all eight million points), his calculations
had taken one-sixth of a second, which mimicked a few hours of actual
oil flow. one of America's 20 Grand Challenges
had been solved - by a
Nigerian who dropped out of school at age 14!
"When he told me the results, I thought he had made a mistake,"
says William Martin, Director of the University of Michigan's
Laboratory for Scientific Computation where Emeagwali spent 13
hours a day for eight months working on the project.
It was enough to win him the Gordon Bell Prize.
Said the Gordon Bell Committee: "The amount of money at stake
is staggering. For example, you can typically expect to recover
10% of a field's oil. If you can improve your production schedule
to get just 1% more oil, you will increase your yield by $400m (at
$20 per barrel in a 20-billion-barrel field)".
Emeagwali's breakthrough made this possible. It changed the view
of a sceptical petroleum industry that massively parallel computers could be
used to recover more oil. Today, seven years on, 10% of all massively
parallel computers are purchased by the oil industry alone. And
the speed of Emeagwali's calculations are helping the OPEC nations
to extract more oil and increase their oil revenue.
Since the breakthrough, Emeagwali has gone from strength to strength.
In March this year, he was adjudged the Pioneer of the Year by the
National Society of Black Engineers in the US. The award was presented
to him at a ceremony in Nashville, Tennesse, attended by 6,000
international scientists and engineers. It is the most prestigious
recognition bestowed by the Society on individuals whose "intellectual
contributions have benefitted all mankind."
Emeagwali was cited for his "discoveries and inventions that led to
the acceptance of massively parallel computing technology, and
in particular their use in the petroleum industry to recover more oil."
The award greatly pleased Emeagwali as he got it ahead of other eminent
scientists and engineers like the US-based Ghanaian Dr Thomas
Mensah who is currently leading the design of advanced laser-guided
weapons (also known as Smart Weapons) such as the US Patriot missiles
which became famous during the Gulf War.
Dr Mensah holds seven patents in fiber optics - three
in fiber optics guided missile technolgy, and four
in manufacturing fiber optics
inexpensively which, experts say, will impact hugely on the
use of fax machines, electronic banking and other
communications gadgetry.
Despite the fame, Philip Emeagwali is not resting on his oars.
This April, he presented a 819-page description of a revolutionary
hyperball computer to the US Office of Patents and Trademarks. His
new computer will enable scientists to better understand the
long-term effects of greenhouse gases and global warming.
He based the
new computer on the "theory of tessellated
models" which he formulated himself. This theory demonstrates
that tessellations (sphere packings) and fast parallel
computations are related fields.
Parallel computers in use today are based on the hypercube or
hypertree networks which do not map efficiently over the
Earth's surface. Emeagwali's hyperball is different. It
was inspired by the complex geometries and patterns he
observed in nature - the arrangements of molecules in crystals
and the structure of bee's honeycombs.
His invention divides the Earth's atmosphere into sections
that simulate a bee's honeycomb. Bees, he says, use the most
efficient method to construct their honeycombs, so a computer
network which emulated the honeycomb will be more efficient and
powerful in forecasting the world's weather patterns a century
or more ahead of time.
"At the moment, there is no computer that can do this. My hyperball
will make this possible," he says.
Married
to an African-American microbiologist, Dale Brown,
who has borne him a 6-year-old son, Emeagwali is
enjoying his days in St Paul, Minnesota. He is an
Internet freak - his personal Website
(http://emeagwali.com) is so hot that it is
accessed by 800 people a day from all over the world.
"They like to know more about my work, and I do my
best not to disappoint
them, he says. "But I would be happier when Africa
finally gets connected to the
information superhighway through the proposed
Africa ONE project. I can't wait
for that day."
Reported in the July/August 1996 issue of New African (London, England).
Click on emeagwali.com for more information.
|