Approaching Infinite Computing

[this webpage uses 5-digit years]

The following are notes for a talk given on Exascale Day 02012 at Scottsdale Community College. [Note: I self-declared 10/18 to be Exascale Day at SCC after learning the NCSA (National Center for Supercomputing Applications at the University of Illinois) proclaimed 10/15 as Petascale Day. I would have scheduled this talk for Petascale Day, but my schedule was a problem.]

The top supercomputer on the June 02012 TOP500 Supercomputing Sites List was Sequoia benchmarked at 16.32 petaflops (16,324,750,000,000,000 flops [floating-point operations per second]). Sequoia is installed at the Department of Energy's Lawrence Livermore National Laboratory. [Japan's K computer was second at 10.510 petaflops followed by Argonne National Laboratory's Mira supercomputer at 8.162 petaflops.]

[update::02013.01.13] The top supercomputer on the November 02012 TOP500 Supercomputing Sites List was Titan at 17.59 petaflops. Titan is at Oak Ridge National Laboratory in Oak Ridge, Tennessee.

On 1 April 02009 (3 1/2 years ago; [update: 3 years 10 months ago on 19 Feb 02013]), I gave a talk titled 20 Petaflops by 02012. The next TOP500 List will be published during November of 02012. The following is from the November 02008 TOP500 List.

   "The Los Alamos system, nicknamed Roadrunner, was slightly 
    enhanced since June and narrowly fended off a challenge by 
    the Cray XT5 supercomputer at Oak Ridge National Laboratory 
    called Jaguar. The system, only the second to break the 
    petaflops barrier, posted a top performance of 1.059 petaflops."
   16.32   petaflops Jun 02012  ( 1.55x in 0.5 years; 55% increase)
   10.51   petaflops Nov 02011  (10.2x in 3.5 years; 942% increase)
    1.026  petaflops Jun 02008  ( 7.5x in 3 years; 650% increase)
    0.1368 petaflops Jun 02005  (i.e. 136.8 teraflops)
    teraflops 01996; gigaflops 01985; megaflops 01964

   ... flops written as whole numbers ...

   16,320,000,000,000,000   (Jun 02012)
   10,510,000,000,000,000   (Nov 02011)
    1,026,000,000,000,000   (Jun 02008)
      136,800,000,000,000   (Jun 02005)

My own calculations...

   petaflops    when
   ===================
     10       02011.00  <-- 10 petaflops (November 02011)
     20       02012.25
     40       02013.50
     80       02014.75
    160       02016.00
    320       02017.25
    640       02018.50
   1280       02019.75  <-- 1.28 exaflops

   1.28 exaflops is 1,280,000,000,000,000,000 flops

Moore's Law growth factors range from 18 to 36 months; however, my calculations show petaflops doubling every 15 months and getting us to exaflops during 4th-quarter of 02019.

When it comes to thinking about today's computing world, it helps to think seconds, not hours; minutes, not days; hours, not weeks; and so on. Think seconds that are approaching zero: nanoseconds, picoseconds, femtoseconds, attoseconds, and so on.

What is Infinite Computing? Infinite performance, infinite storage, infinite bandwidth and infinite visualization.

Converting stuff into bits[1] enables the stuff to be processed by Infinite Computing. And these days more and more stuff is getting is getting converted into bits.

An Infinite Computing environment enables software to accept more inputs and process vastly more variables. And the variables can store smaller small numbers and larger large numbers. [variable domains become -∞ < x < ∞] Infinite storage implies no delete functions. High-speed Internet connections implies access to "Big Data."[2] The outputs from programs can be piped into high performance visualization systems.

Infinite Computing enables real-time to approach real-time.

It is possible that the productivity of STEMers (Scientists, Technologists, Engineers, Mathematicians) might increase by orders of magnitute because of faster execution times; more complex equations; calculations using smaller and larger numbers; and, the visualization of what happens when data approaches zero and/or infinity. STEMers can explore an almost infinite number of "what if" scenarios. Brute force algorithms can be used; therefore, reducing system complexity with respect to design, programming, testing, quality assurance, maintaining, and so on.

A scientist submits a job to the computer.

   in the past...
      + job costs per execution were high
      + jobs took hours to execute
      + number of variables were limited (small data)
      + variable domains were limited
      + brute force algorthims were not an option
      + outputs were difficult to process
      + failures were expensive

   given Infinite Computing...
      + job costs per execution are low
      + jobs take seconds to execute
      + number of variables are almost unlimited (big data)
      + variable domains are almost unlimited
      + brute force algorithms can be designed and implemented
      + outputs can be processed by powerful visualization systems
      + failures are cheap

Informatics [3] is the morphing of data into information. Infinite Computing enables the World Wide Web to become a semantic web, which in turn increases our Informatics. [infinite loop]

What do scientists say when you give them a petaflops supercomputer? ... We need more flops.

Previous Computing Talks (oldest-to-newest)
Notes

[1] Bits are binary digits (i.e. zero and ones). Computers use the binary number system (i.e. base-2 number system).

[2] "Big Data" was a popular buzz-phrase on 10/18/02012. Processing big data requires fast and accurate high-precision number crunching.

[3] "Analytics" was becoming a popular form of Informatics on 10/18/02012. Computers have been doing informatics for a long time. In the past, Informatics was known as data processing and information processing.


Creator: Gerald Thurman [gthurman@gmail.com]
Created: 10 October 02012
Last Modified: Saturday, 05-Jan-2013 10:45:34 MST


More...

Blog.StephenWolfram.com::Latest Perspectives on the Computation Age [02012.10.11]

Begins with...

   "Lots of things have happened in the world in the past 100 years. 
    But I think in the long view of history one thing will end up 
    standing out among all others: this has been the century when 
    the idea of computation emerged." -- Stephen Wolfram

Ends with...

   "I want to leave you with the thought that even after everything 
    that's happened with computers over the past 50 years, we haven't 
    seen anything yet. Computation is a much stronger concept--and 
    actually my guess is it's going to be the defining concept for 
    much of the future of human history." -- Stephen Wolfram