Inventing the World’s Fastest Computer | Philip Emeagwali | The Most Famous Computer Programmers


TIME magazine called him
“the unsung hero behind the Internet.” CNN called him “A Father of the Internet.”
President Bill Clinton called him “one of the great minds of the Information
Age.” He has been voted history’s greatest scientist
of African descent. He is Philip Emeagwali.
He is coming to Trinidad and Tobago to launch the 2008 Kwame Ture lecture series
on Sunday June 8 at the JFK [John F. Kennedy] auditorium
UWI [The University of the West Indies] Saint Augustine 5 p.m.
The Emancipation Support Committee invites you to come and hear this inspirational
mind address the theme:
“Crossing New Frontiers to Conquer Today’s Challenges.”
This lecture is one you cannot afford to miss. Admission is free.
So be there on Sunday June 8 5 p.m.
at the JFK auditorium UWI St. Augustine. [Wild applause and cheering for 22 seconds] [Philip Emeagwali: Contributions to Physics] My contributions to science
stood out for one reason, namely, I worked alone.
For that reason, I won the top prize in supercomputing alone,
and won it back in 1989 and at age thirty-five.
At that time, supercomputing was multidisciplinary.
Supercomputing was a subject that drew heavily from mathematics, physics,
and computer science. Therefore, it was typical
for a team of fifty seasoned research scientists from different fields
to cooperatively work together to solve a grand challenge
initial-boundary problem of mathematical physics,
such as developing the extreme-scaled and the parallel processed
general circulation model or the petroleum reservoir simulator.
In a supercomputing research team, each scientist drew upon
his or her disciplinary knowledge from his or her sub-specialties
within physics, mathematics, and computer science
and with the team integrating its scientific knowledge
and its mathematical techniques from those sub-disciplines.
Some made the egregious mistake of comparing my contributions
to the development of the supercomputer that won the top prize
in supercomputing and did so back in 1989
to the contributions of a team of fifty
supercomputer scientists that recently won the top prize
in supercomputing. It’s like comparing a fifty-person
relay race that covered a total distance
of fifty miles to a one-person race
that covered the same distance of fifty miles.
But more importantly, the recent contributions of the team
of fifty supercomputer scientists was to reconfirm
for the millionth time my primordial discovery
of practical parallel supercomputing that occurred on the Fourth of July 1989.
Today and in China or United States or Japan or the European Union,
a multi-disciplinary team of one thousand supercomputer scientists
might be given a massively parallel supercomputer
that costs up to 1.25 billion dollars. The fastest supercomputer in the world costs
more than the space craft that took men to the moon
and costs more than the budget of each of the forty poorest nations
in the world. The fastest supercomputer in the world
is knowledge-intensive and for that reason
cannot be within the confines of any academic institutions,
in part because, the fastest supercomputer
costs more than the annual budget of any institution of higher education
but, more importantly, the body of knowledge
of parallel supercomputing is extraordinarily deep
and too expansive to reside in its entirety
within any single academic institution. [ Sometimes, the Impossible Remains Impossible] When parallel processing meets
the biggest questions in computational science,
the impossible-to-solve becomes possible-to-solve.
Parallel processing is the vital technology
that enables us to ask the biggest questions
and then find new answers to those previously unanswered questions.
I began my quest for practical parallel supercomputing
in the realm of science fiction, namely, by imagining
the one binary million email wires that I must use to ingest my data
and do so from outside a new internet that is a new global network of commodity-off-the-shelf
processors. I began that quest
by imagining how to move my data efficiently
and move them inside my sixty-four binary thousand processors
that were tightly-coupled to each other and that shared nothing
between each other. I also imagined my email messages
as sent to and received from two-raised-to-power sixty-four
processors that I imagined as encircling a globe
in sixty-four dimensional hyperspace. I discovered that
our post human descendants of Year Million
will forever find it impossible to construct their parallel supercomputer
that has a one processor to one vertex correspondence
with the two-raised-to-power sixty-four vertices of the cube
in sixty-four dimensional hyperspace. [The World’s Fastest Computer] [The Limit of the Supercomputer] Today’s grand challenge questions
are more complex than ever. An example of a grand challenge problem
is how to massively parallel process the extreme-scale
computational fluid dynamics codes that must be parallel executed
when modeling the flow of blood through the human
cardiovascular system. Parallel processing
is an entirely new approach to modern computer science.
Yet, there is a practical limit to the theoretically unlimited
speed of the parallel supercomputer. Often, there is a limit
to what seemed unlimited. A story about the origin of chess
contains an important lesson on why the parallel supercomputer
cannot be constructed with a processor-to-vertex correspondence
with the hypercube in the sixty-fourth dimensional hyperspace. About 800 years ago,
King Shirham of India loved to play games.
Eventually, the King mastered and became bored with all the games
known to games masters. The King invited Sissa Ben Dahir,
his Grand Vizier, or prime minister, to his palace
and commanded Sissa Ben Dahir to invent
the toughest game in the world. After a year of meditation
and hard work, the Grand Vizier, returned to the palace. “Have you invented the toughest game
in the world?” the King asked his Grand Vizier.
“Yes,” the Grand Vizier answered. “I call it ‘Chaturanga.’”
That new game, Chaturanga, is the precursor of the modern chess,
a game that beckons upon the most intelligent persons.
Chaturanga is played on a game board
that is comprised of eight rows and eight columns,
or sixty-four, black and white checkered squares.
After playing Chaturanga the King exclaimed.
“This is the toughest game in the world! Name your reward for this invention.”
The Grand Vizier thought carefully and then said:
“My reward for inventing Chaturanga is a pile of rice.”
“Why don’t you ask for gold, instead of rice,”
the King wondered aloud. Gesturing to the sixty-four squares
on his new chessboard, Chaturanga, the Grand Vizier asked for
a grain of rice for the first square, two grains for the second,
and four grains for the third. The King thought this was a silly request.
“Is that all? Seven grains.” the King interrupted the Grand Vizier.
“No,” the Grand Vizier continued “each square got double
what the last square got. All sixty-four squares
get their grains of rice.” Puzzled, the King protested.
“I have a greater reward: Take my daughter’s hand in marriage.”
“Your majesty,” the Vizier said, “I’m a happily married man.”
“Oh, how can I forget Mrs. Vizier?” The King commanded his aides
to bring a spoon full of rice and fill all sixty-four squares
of the new chessboard, with each square allocated
double what the previous square got.
The servants started counting the grains of rice.
One, two, four, and soon the first teaspoon of rice
was used up. Then four bags of rice were required, then
the King got pale as 1,024 bags of rice,
then one binary million bags of rice
were gone at the twentieth square. The amount rice needed
seemed infinite! The total grains of rice needed
for all sixty-four squares was two-raised-to-power sixty-four minus one,
or 18 quintillion grains of rice, or about eighteen
followed by eighteen zeroes grains of rice.
The total bags of rice the Grand Vizier demanded
was equal to the total amount of rice that was ever harvested
by all the rice farmers that ever lived on planet Earth.
That amount of rice will cover the Earth,
many times over and has the mass of Mount Everest. [Supercomputing Lessons Learned from Chess] I shared this story
from the 13th century India and did so to highlight
the invisible limits to the speed of the future
planetary-sized supercomputer. But I also use this story
to explain to children why it’s impossible
to scrub off a picture that’s gone viral on the internet.
You share your picture with two facebook friends,
they share it with four friends, and when your picture has gone viral
you cannot scrub it out of the Internet. I also shared this story because
I once speculated that supercomputer scientists
of the 22nd century, or farther, could parallel process across
their internet that could be defined and outlined
by two-raised-to-power sixty-four commodity processors
that had a one-to-one processor-to-vertex correspondence
and had that relationship with the vertices of a cube
that is tightly-circumscribed by a globe in the sixty-fourth dimensional
hyperspace. [The World’s Fastest Computer] I’m Philip Emeagwali.
I’m known as the first massively parallel supercomputer scientist
and as the first person to discover
how a million processors can be fused together
by as many email messages and fused together
to form one whole cohesive unit that is a new computing machinery
that is the world’s fastest computer that made the news headlines
back in 1989. The new supercomputer
that is a new internet that I invented on July 4, 1989
is radically different from the constituent processors
that it originated from. As an aside and in my dictionary,
the words “computer” and “internet” are like two sides of the same coin.
I believe that the 22nd century supercomputer scientist
cannot have an internet that is not also a parallel supercomputer,
or vice-versa. The moral of my story
of the origin of chess in the 13th century India
and its lessons for the planetary-sized
supercomputer-hopeful of the 22nd century
was that at the beginning I was as unaware as the King
but at the end I became as knowledgeable
as the Grand Vizier. At first and back in the early 1980s,
I grossly underestimated the power of doubling.
I once thought that I could simultaneously program
two-raised-to-power sixty-four number of processors
and synchronously send and receive email messages across
a new internet that I visualized as
a new global network of sixty-four times
two-raised-to-power sixty-four number of bi-directional email wires.
A third important lesson of my story of the origin of chess
in the 13th century India lies in the uncontrolled growth
of the population of Nigeria, my country of birth.
Nigeria is now the 7th most populous nation in the world.
Nigeria is a little bigger than Texas but could grow, by mid-21st century,
to become the third most populous nation in the world.
Last year, Nigeria welcomed 5.5 million newborns,
or 40 percent more babies than the United States.
This year, Nigeria will welcome more than the population of Libya
or Norway or New Zealand. The planetary-sized vision
that inspired my contribution to the development of the supercomputer
is this: The Earth is enshrouded by fluids,
namely, the atmosphere and the oceans as well as the rivers and lakes
and also surrounded by subterranean fluids
such as crude oil, natural gas, and water. To invent a new internet
that is a new supercomputer de facto, I visualized 65,536 processors
as equal distances apart and within the fluids
that enshroud the Earth. I envisioned a globe
in the sixteenth dimension that was encircled by
two-raised-to-power sixteen, or 65,536, commodity-off-the-shelf processors
that were distributed equal distances apart from each other
and that were tightly-coupled to each other
and that shared nothing between each other.
I envisioned each processor as having its own operating system.
I envisioned those 64 binary thousand processors as embedded within the fluid,
or atmospheres and oceans, that enshroud the Earth, or globe.
I envisioned that globe as tightly encircling a cube
with sixteen times two-raised-to-power sixteen,
or 1,048,576, bi-directional edges. In the modern configuration
of supercomputers and at one foot per email wire,
those email wires will total 200 miles of cables.
Each vertex on the surface of that globe
was my metaphor for one processor. Each bi-directional edge
on the surface of that globe was my metaphor for one email wire.
That globe is my metaphor for the Earth.
That new global network of sixty-four binary thousand processors
is one of the two internets that I invented as two supercomputers
and is the reason I was profiled in books
such as the one titled: “History of the Internet.”
I envisioned each processor as simulating the motions
of the nearest three thousand square miles of fluids. [My Contributions to Algebra] I invented a new supercomputer
that encircled the globe in the way the internet does,
and that could be used to solve never-before-solved problems in algebra. I
invented two new supercomputers. My first supercomputer
was constructively reduced to practice as an ensemble of processors
that encircled the globe in the way the Internet does. My second supercomputer
was an actual reduction to practice of 65,536 processors
that encircled the globe in the sixteenth dimension
and did so in the way the Internet does. That parallel supercomputer
became super by computing faster than any
scalar or vector supercomputer. That new supercomputer
enables the computational mathematician
and physicist to answer previously unanswerable questions
arising in extreme-scale algebra. Such questions are recurring decimals
in the grand challenge problems of supercomputing. By definition, algebra
is the generalization of arithmetic. In high school algebra, two letters represent
two numbers. In the supercomputer algebra
that arises from trying to discover and recover
otherwise elusive crude oil and natural gas, one trillion letters
must represent one trillion numbers arising from a system of
one trillion equations of algebra. Those one trillion equations
are evenly distributed across one million processors that, in turn, solves
them in parallel by computing one million calculations
at once. Thank you. I’m Philip Emeagwali. [Wild applause and cheering for 17 seconds] Insightful and brilliant lecture

Leave a Reply

Your email address will not be published. Required fields are marked *