[LINK] Chinese Supercomputer

[Tom Koltai]

1.27 and 1.75 Petaflops.

Interestingly dull.

I have it on a most excellent authority that Australia is doing a little
better in the Super computer department than both of these examples.

Glad to see we're out-doing someone, somewhere in something IT
based.....

And no, sorry to all, can't actually say who the excellent authority is.
But it isn't .gov or .edu.

Most interesting times..... Continued below....

> -----Original Message-----
> From: link-bounces at mailman1.anu.edu.au 
> [mailto:link-bounces at mailman1.anu.edu.au] On Behalf Of 
> stephen at melbpc.org.au
> Sent: Tuesday, 1 June 2010 9:58 PM
> To: link at anu.edu.au
> Subject: [LINK] Chinese Supercomputer
> 
> 
> Chinese supercomputer ranked second-fastest
> 
> 1st June 2010 
> <http://www.chinadaily.com.cn/bizchina/2010->
06/01/content_9918532.htm>
> 
> 
> A Chinese supercomputer has ascended to the second place in 
> the world in 
> terms of speed .. 
> 
> The Dawning Nebulae of the Shenzhen-based National 
> Supercomputing Center 
> has reached a sustained computing speed of 1.27 petaflops, 
> The Cray Jaguar supercomputer, which is based at the Oak 
> Ridge National 
> Laboratory in Tennessee, is still the fastest computer in the 
> world, with 
> a speed of 1.75 petaflops. The data was measured in last November. 

(The 1.75 is for short term and not extended peak load - the reality is
that the Cray is locked at 1.5 Pf's.)
Ummm, no it isnt. I understand the new top speed is over 2.41 Pf or 10
to the 15.25

> Previously, a supercomputer at the Tianjin-based National 
> Supercomputing 
> Center gained the best-ever ranking, the fifth, but now it is at the 
> seventh place. 
> 
> The United States, which has 282 of the world's fastest 500 
> computers on 
> the new list, remains a leader in this area. 

Speed increases past these mind-numbing numbers would seem to depend
more on properties of certain metals not fracturing under extreme liquid
cooling conditions.

Which I guess means that we must be on the verge of developing that cold
fuse and ion thingy.

Although the AU version uses a lot less power than the Cray ergo
substantially less cooling.

In my lifetime, we have gone from a few thousand cycles to .....
And from 8 bit to 128 bit.......

Boy I would like to be around in fifty years.....

According to moores law, we should by then all be running around with
terraflop earrings that connect us to the community mind meld.

I have just one question.... Where are we going to fit the 1 tonne (5
HP) aircon unit to cool the earrings ?

Not sure if Linkers realise it, but even Cesium needs to be chilled to 0
K to ensure an alsmost accurate time freq measurement... Almost accurate
? Well.. I quote "The second is the duration of 9 192 631 770  periods
of the radiation corresponding to the transition between the two
hyperfine levels of the ground state of the caesium 133 atom.
It follows that the hyperfine splitting in the ground state of the
caesium 133 atom is exactly 9 192 631 770 hertz, nu(hfs Cs) = 9 192 631
770 Hz. This definition refers to a caesium atom at rest at a
temperature of 0 K."

On that basis what chance have we mere mortals on getting our hands on
128 bit Venus processors and not cooking them five seconds later ?

Sigh....