7 ms·

Is The Chinese Supercomputer Our Sputnik?

14 points by tmckd 16y ago

nkassis 16y ago
There was an article posted here on HN not too long ago about Japan's 5th generation computer project. It was more about what killed logic programming but this China+supercomputing thing reminds me of this a little. Here's the article: http://vanemden.wordpress.com/2010/08/21/who-killed-prolog/ http://vanemden.wordpress.com/2010/08/21/who-killed-prolog/ EDIT: Wikipedia page about the project: http://en.wikipedia.org/wiki/Fifth_generation_computer http://en.wikipedia.org/wiki/Fifth_generation_computer
thelema314 16y ago
The Chinese CPU is junk, and the chinese supercomputer is built out of US parts - I wouldn't be too apocalyptic about technological superiority for some years yet.
- jlouis 16y ago
  The CPUs are just a commodity. The interconnect is where it is at.
  - heresy 16y ago
    Until the interconnect is a commodity, and at any point the advances they've made there can be as easily utilized in building new US supercomputers. I don't really see the point of this chest thumping, whoever does it, it just seems like so much posturing. It's fast. Great, but get back to actual science please.
dotBen 16y ago
Asks an American tax payer: Why does America need to have the fastest super-computer? (the premis of this and other coverage of this topic)
- dmlorenzetti 16y ago
  One could flip that question around and ask: Why does China need the fastest machine? The significance of the story isn't that America no longer ranks first. I believe that in general, the really fast machines serve one of two purposes-- simulating nukes (for example, to test new designs), and simulating global environment (for example, to improve climate predictions).
- ihumanable 16y ago
  I think the article is making the argument that the goal of having the fastest X or the biggest X or being the first to X is not the point. Supercomputers have real value, in the sense that they are used to solve real problems that can't be solved otherwise. The bigger point seems to be that the benefit in getting to X is that you normally create a lot of technology that ends up being beneficial to the economy in general. Having an uninterruptable communication network led to the internet. Being able to deploy ICBMs across the nation led to the Highway System. Going to space created a ton of new materials and manufacturing systems that have created products that couldn't have existed before. Having the fastest supercomputer is the trees, the forest is the wealth of technology and industries born out of that race.
  - dotBen 16y ago
    Supercomputers have real value, in the sense that they are used to solve real problems that can't be solved otherwise Sure, and this answers the question "why have super-computers", indeed, the US has several. I'm not just sure if any of the reasons you specify, all valid, suddenly become invalid when there is simply a delta between the power of US vs China systems.
CamperBob 16y ago
No. That'll happen when the Chinese reach the Moon.
- ximeng 16y ago
  In 2013: http://en.wikipedia.org/wiki/Chang%27e_3 http://en.wikipedia.org/wiki/Chang%27e_3
Tichy 16y ago
"Supercomputing is at the basis of many important scientific and commercial enterprises in the United States, and it is very important that we not relinquish our lead in this area." Details, please. Otherwise it is just blubbering.
- possibilistic 16y ago
  I'm studying to become an interdiciplinary computational biochemist, so perhaps I can lend some perspective. Supercomputers are increasingly used in the sciences to scale down the search space of massively complicated problems such as protein folding, molecular docking, and chemical synthesis. In several instances they lead us to novel, non-intuitive results, sometimes as small as finding a functional group prefering a gauched vs anti conformation that changes a particular interaction completely. With drug design specifically, we can cut down the chemical search space necessary to find inhibitor molecules and search for potential problematic harmful interactions before ever running in vitro assays. Science today is more serendipity and hard work than acts of sheer brilliance. Having high-throughput tools that save us from looking too hard give us more time to think and more quickly update our models. Supercomputing will lead us to smarter medicine, cures for cancer, better materials, and a more energy-efficient planet. Computer science is a major cornerstone of today's scientific endeavors, and we increasingly need to encourage interest and investment in this field.
fhe 16y ago
this reads like what the Japanese did in WWII: racing to build the largest battleship when naval warfare had shifted to carriers and airplanes (http://en.wikipedia.org/wiki/Japanese_battleship_Yamato http://en.wikipedia.org/wiki/Japanese_battleship_Yamato).
- InclinedPlane 16y ago
  To be fair, they also built a crap-ton of carriers and airplanes.
  - hga 16y ago
    To be even more fair, so did we, since neither of us knew ahead of time which or what mix would prove decisive.
adaml_623 16y ago
Such a sad US vs THEM post. Americans are just going to have to accept that the playing field is becoming more level and eventually a country with 1 Billion people is going to be more productive than a country with 300 Million people. It's just arithmetic. And it's not the end of the world. It's ok to just live in a country and have a cool life.
- gaius 16y ago
  How do you figure that? The size of a country's population is nothing to do with anything, really. History is full of examples.
  - garply 16y ago
    That's good to point out, but China is consciously adopting a Western legal and economic system (with some notable modifications) on which to run its society. If they do it successfully, the larger amount of people should give the country a greater amount of productivity than the US.
    - gaius 16y ago
      That's a very fine balancing act. Everything is magnified. What if 10% unemployment in the US economy at an equivalent tech level translates to 60% in China? Perhaps the Chinese will choose to keep productivity at a controlled level to prevent social unrest.
  - adaml_623 16y ago
    What no that's a crazy comment. Expand on your one line with your examples. History is full of examples of countries with big populations doing big things. And countries with tiny populations achieving less. Of course the number of people makes a difference.
    - gaius 16y ago
      Well on the subject of China let's start with Ghengis Khan conquering not only China but a vast empire... Or Athens, or Rome (these pre-date Greece and Italy existing as countries)... Or England, Holland and Portugal building vast trading empires and driving the Renaissance and the Industrial Revolution. Numbers are nothing. Culture is fate.
jlouis 16y ago
The article is wrong on " The fastest supercomputer has, since 1994, resided in the United States at the Oak Ridge National Laboratory, where it is used for scientific computing, among other purposes." I present to the thee, the Japanese Earth Simulator: http://en.wikipedia.org/wiki/Earth_Simulator http://en.wikipedia.org/wiki/Earth_Simulator

14 points by tmckd 16y ago

There was an article posted here on HN not too long ago about Japan's 5th generation computer project. It was more about what killed logic programming but this China+supercomputing thing reminds me of this a little. Here's the article: http://vanemden.wordpress.com/2010/08/21/who-killed-prolog/ http://vanemden.wordpress.com/2010/08/21/who-killed-prolog/ EDIT: Wikipedia page about the project: http://en.wikipedia.org/wiki/Fifth_generation_computer http://en.wikipedia.org/wiki/Fifth_generation_computer

thelema314 16y ago

The Chinese CPU is junk, and the chinese supercomputer is built out of US parts - I wouldn't be too apocalyptic about technological superiority for some years yet.

jlouis 16y ago
The CPUs are just a commodity. The interconnect is where it is at.
- heresy 16y ago
  Until the interconnect is a commodity, and at any point the advances they've made there can be as easily utilized in building new US supercomputers. I don't really see the point of this chest thumping, whoever does it, it just seems like so much posturing. It's fast. Great, but get back to actual science please.

dotBen 16y ago

Asks an American tax payer: Why does America need to have the fastest super-computer? (the premis of this and other coverage of this topic)

dmlorenzetti 16y ago
One could flip that question around and ask: Why does China need the fastest machine? The significance of the story isn't that America no longer ranks first. I believe that in general, the really fast machines serve one of two purposes-- simulating nukes (for example, to test new designs), and simulating global environment (for example, to improve climate predictions).
ihumanable 16y ago
I think the article is making the argument that the goal of having the fastest X or the biggest X or being the first to X is not the point. Supercomputers have real value, in the sense that they are used to solve real problems that can't be solved otherwise. The bigger point seems to be that the benefit in getting to X is that you normally create a lot of technology that ends up being beneficial to the economy in general. Having an uninterruptable communication network led to the internet. Being able to deploy ICBMs across the nation led to the Highway System. Going to space created a ton of new materials and manufacturing systems that have created products that couldn't have existed before. Having the fastest supercomputer is the trees, the forest is the wealth of technology and industries born out of that race.
- dotBen 16y ago
  Supercomputers have real value, in the sense that they are used to solve real problems that can't be solved otherwise Sure, and this answers the question "why have super-computers", indeed, the US has several. I'm not just sure if any of the reasons you specify, all valid, suddenly become invalid when there is simply a delta between the power of US vs China systems.

CamperBob 16y ago

No. That'll happen when the Chinese reach the Moon.

ximeng 16y ago
In 2013: http://en.wikipedia.org/wiki/Chang%27e_3 http://en.wikipedia.org/wiki/Chang%27e_3

Tichy 16y ago

"Supercomputing is at the basis of many important scientific and commercial enterprises in the United States, and it is very important that we not relinquish our lead in this area." Details, please. Otherwise it is just blubbering.

possibilistic 16y ago
I'm studying to become an interdiciplinary computational biochemist, so perhaps I can lend some perspective. Supercomputers are increasingly used in the sciences to scale down the search space of massively complicated problems such as protein folding, molecular docking, and chemical synthesis. In several instances they lead us to novel, non-intuitive results, sometimes as small as finding a functional group prefering a gauched vs anti conformation that changes a particular interaction completely. With drug design specifically, we can cut down the chemical search space necessary to find inhibitor molecules and search for potential problematic harmful interactions before ever running in vitro assays. Science today is more serendipity and hard work than acts of sheer brilliance. Having high-throughput tools that save us from looking too hard give us more time to think and more quickly update our models. Supercomputing will lead us to smarter medicine, cures for cancer, better materials, and a more energy-efficient planet. Computer science is a major cornerstone of today's scientific endeavors, and we increasingly need to encourage interest and investment in this field.

fhe 16y ago

this reads like what the Japanese did in WWII: racing to build the largest battleship when naval warfare had shifted to carriers and airplanes (http://en.wikipedia.org/wiki/Japanese_battleship_Yamato http://en.wikipedia.org/wiki/Japanese_battleship_Yamato).

InclinedPlane 16y ago
To be fair, they also built a crap-ton of carriers and airplanes.
- hga 16y ago
  To be even more fair, so did we, since neither of us knew ahead of time which or what mix would prove decisive.

adaml_623 16y ago

Such a sad US vs THEM post. Americans are just going to have to accept that the playing field is becoming more level and eventually a country with 1 Billion people is going to be more productive than a country with 300 Million people. It's just arithmetic. And it's not the end of the world. It's ok to just live in a country and have a cool life.

gaius 16y ago
How do you figure that? The size of a country's population is nothing to do with anything, really. History is full of examples.
- garply 16y ago
  That's good to point out, but China is consciously adopting a Western legal and economic system (with some notable modifications) on which to run its society. If they do it successfully, the larger amount of people should give the country a greater amount of productivity than the US.
  - gaius 16y ago
    That's a very fine balancing act. Everything is magnified. What if 10% unemployment in the US economy at an equivalent tech level translates to 60% in China? Perhaps the Chinese will choose to keep productivity at a controlled level to prevent social unrest.
- adaml_623 16y ago
  What no that's a crazy comment. Expand on your one line with your examples. History is full of examples of countries with big populations doing big things. And countries with tiny populations achieving less. Of course the number of people makes a difference.
  - gaius 16y ago
    Well on the subject of China let's start with Ghengis Khan conquering not only China but a vast empire... Or Athens, or Rome (these pre-date Greece and Italy existing as countries)... Or England, Holland and Portugal building vast trading empires and driving the Renaissance and the Industrial Revolution. Numbers are nothing. Culture is fate.

jlouis 16y ago

The article is wrong on " The fastest supercomputer has, since 1994, resided in the United States at the Oak Ridge National Laboratory, where it is used for scientific computing, among other purposes." I present to the thee, the Japanese Earth Simulator: http://en.wikipedia.org/wiki/Earth_Simulator http://en.wikipedia.org/wiki/Earth_Simulator