Approaching Infinite Computing; 10^18 in 2018?

The number one supercomputer on the November 2012 TOP500 List was benchmarked at 17.59 petaflops. [Titan at ORNL] The computing road map predicts exascale computing by the end of this decade (i.e. 02018-02020).

17.59 petaflops = 17,590,000,000,000,000 flops
[peta- metric prefix = 10^15 (1 quadrillion)]
[flops floating-point operations per second]

[exa- metric prefix = 10^18 (1 quintillion)]
1 exaflops = 1,000,000,000,000,000,000 flops
[exaflops is a billion billion flops]

The June 02008 TOP500 List was the first time a petaflops computer was number on the list. [RoadRunner at 1.026 petaflops]


Moore's Law

Moore's Law in a nutshell: The number of transistors on a microchip will double every two years. [Gordon E. Moore circa 01965]


With Infinite Computing...

Cost per one gigaflops: $8.3 trillion in 01961; 73 cents ($0.73) in August 02012. [source: Wikipedia.org]

...fewer space & time tradeoffs...
...more variables (inputs, outputs, intermediate results)...
...more advanced mathematical expressions/equations...
...variables have larger domains and ranges [1]...
...more what if scenarios can be tested...
...delete becomes obsolete...
...more chances to use brute force...
   Brute force can result in simpler algorithms.
   Simpler algorithms are simpler to program.
   Simpler code can result is "better" software.

What does a STEMer say when you give them a petaflops computer? I need more flops!


Output generated by a C program executed on a GNU/Linux system on 3 May 02013.

   largest whole number:  18,446,744,073,709,551,615 (18.45 quintillion)

   largest floating-point number: 1.189731 x 104932 (4933 digit number)

I had to make a video about these two numbers...


hpc <=> hpvs (two-way pipe)

What will STEMers do (create, discover, innovate, invent, ...) given high performance computing systems and high performance visualization systems?

Computing jobs (e.g. simulations) that took hours to run are completed in seconds. Costs per run approaching $0.


Wired.com::Science::The End of Theory: The Data Deluge Makes the Scientific Method Obsolete [Chris Anderson; June 2008]

   "Petabytes allow us to say: 'Correlation is enough.' We can stop 
    looking for models. We can analyze the data without hypotheses 
    about what it might show. We can throw the numbers into the 
    biggest computing clusters the world has ever seen and let 
    statistical algorithms find patterns where science cannot."

   "Correlation supersedes causation, and science can advance
    even without coherent models, unified theories, or really
    any mechanistic explanation at all."

   "There's no reason to cling to our old ways. It's time to ask: 
    What can science learn from Google?"

HBR.org::Data Scientist: The Sexiest Job of the 21st Century [Harvard Business Review; October 2012]

"I've sort of become convinced that data science is an important component of general education." -- Stephen Wolfram ["Sergey Brin was an intern at my company long before he started Google."]

YouTube.com::The Computational Knowledge Revolution - Stephen Wolfram [published 3 May 02013]