Up next

Philip Emeagwali | How I Invented a New Internet that is a New Supercomputer

King_Tut - 444 Views
Published on 18 Nov 2018 / In Black History

For complete transcript and lecture series, visit http://emeagwali.com
or https://emeagwali.wordpress.co....m/2018/10/05/how-i-i

For live speaking engagements and tour schedules, call 202-203-8724 or email info@emeagwali.com or http://emeagwali.com.

Scientific knowledge is the first son of God.
Science pre-existed before humanity
and before our planet, the Earth,
was formed 4.6 billion years ago.
Back in 1989,
one of the science news headlines
was that an African Supercomputer Wizard
in the United States
had experimentally discovered
how and why parallel processing
makes computers faster
and makes supercomputers fastest
and invented
how and why to use
that new supercomputer knowledge
to build a new supercomputer
that encircled a globe
and encircled it
in the manner
the internet encircled a globe.
I am that African supercomputer scientist
that was in the news
back in 1989.
I was in the news
for experimentally discovering
that parallel processing
is an entirely new way of supercomputing
across thousands or millions or billions
of processors.
Parallel processing
is defined as the technique
of fastest supercomputing
that is fastest
by computing many things
at once, or in parallel,
instead of computing only one thing
at a time, or in sequence.
Prior to my 1989 experimental discovery,
parallel processing was widely caricatured
and rejected
as a huge waste of everybody’s time.
Parallel processing
was rejected for four reasons.
The first reason
the parallel processing supercomputer
was rejected
was because supercomputing in parallel
had performance problems.
That is, in the 1980s and earlier,
parallel processing supercomputers
could not compute faster than
sequential processing supercomputers.
The second reason
the parallel processing supercomputer
was rejected
was because it was physically impossible
to experimentally discover
how to harness
64 binary thousand processors
and harness them
to compute together to solve
any of the twenty toughest problems
in supercomputing.
Those extreme-scale problems
were called the twenty Grand Challenges
of supercomputing.
The third reason
the parallel processing supercomputer
was rejected
was that programming supercomputers
to solve a system of coupled, nonlinear,
and time-dependent
partial differential equations
of a new calculus
made research computational mathematicians
deeply uncomfortable.
In particular, to parallel process
via emails
sent to and from
sixteen-bit long email addresses
and to parallel process
the most dense, abstract,
and impenetrable equations
and to parallel process
their algebraic approximations
and to parallel process
their floating-point arithmetical calculations
that must be executed across
sixteen times
two-to-power sixteen,
or across one binary million, email wires
is like dancing in the fire.
The fourth reason
the parallel processing supercomputer
was rejected
was that I, its discoverer,
was black and African.
My research and experimental discovery
of parallel processing
was not taken seriously
in the late 1970s and early 1980s.

My 1,057 page research report
on the massively parallel processing supercomputer
was rejected six times
and rejected by three universities
and rejected by scientific journals
before it was eventually accepted
by the supercomputer community.
In the 1980s, the massively
parallel processing supercomputer
was unfathomable
and for that reason
a president of an American university
that had an annual research expenditure
of one billion dollars
and his five supercomputer experts
threw my one thousand
and fifty-seven [1,057]-page
supercomputer research report
into the trash.
When a newspaper journalist
writing about my experimental discovery
came to interview
those five supercomputer experts
they couldn’t do the interview.
The reason was that
they never read or understood
my supercomputer research report.
So I was not taken seriously
until The Computer Society
of the IEEE—The Institute of Electrical
and Electronics Engineers—
gave me the top prize in supercomputing.
To put my dilemma in context,
back in the 1980s,
it was impossible
for an all-white scientific jury
to give me the top award
in computer science.
The award committees
asked for my photograph
or insisted on a face-to-face interview
that will reveal
the fact that I am black and African.
In the 1980s,
only one award committee
did not demand my photograph.
I won that award
and it made the news headlines
that a black African
had won the top prize
in supercomputing.
The controversy
prompted the award committee
to change their rules
and to demand a face-to-face lecture
that, in turn, made it impossible
for other black supercomputer scientists
to win the top prize in supercomputing.

Philip Emeagwali Lecture 180120-1

Show more
0 Comments sort Sort by