FASTER THAN A ’57 CHEVY

November 5, 2016


I grew up in the ’50s in post-world war two (WWII) decade.  It truly was a very simple time as compared to the chaotic, time-obsessed, “hair-on-fire”, get-it-done-at-any cost times we experience today.  One expression I remember very clearly was: “faster than a ’57 Chevy”.  Anything over walking speed was faster than a ’57 Chevy.  This, of course, was handed down from the older kids to guys my age.  The object of that expression may be seen below.

57-chevy

(I told you those were much more simple days.) If we only knew what was coming down the pike, we would have never never used that expression.  You know what is really faster than a ’57 Chevy?  Let’s take a look.

This month the Top 500 biannual ranking of the world’s fastest, publicly known supercomputers will be updated.  The list release will coincide with SC16, the International Conference for High Performance Computing, Networking, Storage and Analysis held in Salt Lake City from November 13 to November 18. The last Top 500 update in June revealed that China maintained its grip on the number one spot with the new and surprising Sunway TaihuLight device, which reached ninety-three (93) petaflops per second or ninety-three quadrillion calculations per second or “faster than a ’57 Chevy”.

Let’s refresh our memory.  A petaflop is a measure of a computer’s processing speed and can be expressed as:

  • A quadrillion (thousand trillion) floating point operations per second (FLOPS)
  • A thousand teraflops
  • 10 to the 15th power FLOPS
  • 2 to the 50th power FLOPS

THE LIST

NUMBER 1: National Supercomputing Center in Wuxi, China: Sunway TaihuLight with 10,649,600 cores running 15,371 Kw.  The Sunway is shown below:

sunway-taihulight

NUMBER 2:  National Super Computer Center in Guangzhou, China: Tianhe-2(MilkyWay-2) with 3,120,000 cores running 17,808 Kw.

tianhe-2

NUMBER 3:  DOE/SC/OAK RIDGE NATIONAL LABORATORY, UNITED STATES:  Titan-Cray XK7: 560,640 cores running 8,209 Kw.

titan-cray-xk7

NUMBER 4:  DOE/NNSA/LLNL, UNITED STATES:  Sequioa-BlueGene/Q:  1,572,864 cores running 7,890 Kw.

sequoia-blue-genie

NUMBER 5:  RIKEN ADVANCED INSTITUTE FOR COMPUTATIONAL SCIENCE, JAPAN: K Computer SPARC64: 705,024 cores running 12,660 Kw.

k-computersparc64

Not only has China outdone itself in terms of the fastest supercomputer, it is now home to the largest number of supercomputers on the list.  One hundred sixty-seven (167) to be exact.  This doubles the number in the United States.  This year marks the first time since the Top 500 rankings began twenty-three (23) years ago, that the United States cannot lay claim to the most machines on the list. All is lost—well not quite.

In September, the DOE’s Exascale Computing Project or ECP, announced the first round of funding for advanced computers.  It awarded $39.8 million to fifteen (15) application development proposals for full funding and seven proposals for seed funding.  This is significant and will provide necessary financing to keep up and even surpass the Chinese.

HISTORY:

If we look at the history of computing power, we see the following:

history-of-computing-power-performance

Compare that with the cost of computing power:

evolution-of-computing-power-costs

CONCLUSIONS: 

As you can see, the advances in computing power are remarkable but at a significant cost.  Speed vs. cost.  This is one very expensive technology.  Fortunately, the computers we mortals use, do not require the speeds cutting-edge technology requires.  With that being the case, computing power on the domestic scale has decreased significantly over the years.

Advertisements

What do you think?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: