[LINK] Death of Moore's Law Will Cause Economic Crisis
stephen at melbpc.org.au
stephen at melbpc.org.au
Wed Mar 23 22:23:29 AEDT 2011
> Gordon Moore's famous law about the doubling of transistor density and
> power every 2 years will not only end it could bring economic disaster
> in its wake, respected scientist Michio Kaku has predicted in a new
> book. [ Link http://ho.io/n3nt ]
> Kaku sets out the crunch moment as being the point at which ultraviolet
> light can no longer tuned to etch ever smaller circuits on to silicon
> wafers, which on current trends will kick in less than a decade from
> now. From that moment on, Moore's Law will gradually diminish, and the
> effects will not only be technological but economic.
Alternatively, one chip with a trillion bytes of memory & 128 processors:
> Subject: [LINK] Processing and Memory
> Date: Fri Mar 4 15:40:19 EST 2011
Remapping Computer Circuitry to Avert Impending Bottlenecks
By JOHN MARKOFF Posted 01 March 2011 - 02:43 PM
PALO ALTO, Calif. Hewlett-Packard researchers have proposed a
fundamental rethinking of the modern computer for the coming era of
nanoelectronics a marriage of memory and computing power that could
drastically limit the energy used by computers.
Today the microprocessor is in the center of the computing universe, and
information is moved, at heavy energy cost, first to be used in
computation and then stored.
The new approach would be to marry processing to memory, to cut down
transportation of data and reduce energy use.
In an article published in IEEE Computer in January, Parthasarathy
Ranganathan, a Hewlett-Packard electrical engineer, offers a radical
alternative to todays computer designs known as exascale processors.
Today, computers constantly shuttle data back and forth among faster and
The systems keep frequently used data close to the processor and then
move it to slower and more permanent storage when it is no longer needed
for the ongoing calculations.
In this approach, the microprocessor is in the center of the computing
universe, but in terms of energy costs, moving the information, first to
be computed upon and then stored, dwarfs the energy used in the actual
"Whats going to be the killer app 10 years from now?" asked Dr.
Ranganathan. "Its fairly clear its going to be about data; thats not
rocket science. In the future every piece of storage on the planet will
come with a built-in computer."
To distinguish the new type of computing from todays designs, he said
that systems will be based on memory chips he calls "nanostores" as
distinct from todays microprocessors.
They will be hybrids, three-dimensional systems in which lower-level
circuits will be based on a nanoelectronic technology called the
memristor, which Hewlett-Packard is developing to store data.
The nanostore chips will have a multistory design, and computing circuits
made with conventional silicon will sit directly on top of the memory to
process the data, with minimal energy costs.
Within seven years or so, experts estimate that one such chip might store
a trillion bytes of memory (about 220 high-definition digital movies) in
addition to containing 128 processors, Dr. Ranganathan wrote.
If these devices become ubiquitous, it would radically reduce the amount
of information that needs to be shuttled back and forth in future data
A 10-petaflop supercomputer scheduled to be built by I.B.M. next year
will consume 15 megawatts of power, roughly the electricity consumed by a
city of 15,000 homes.
An exascale computer, but built with todays microprocessors, would
require 1.6 gigawatts. That would be roughly one and half times the
amount of electricity produced by a nuclear power plant. (snip) ...
The panel did, however, support Dr. Ranganathans memory-centric
approach. It found that the energy cost of a single calculation was about
(A picojoule is one millionth of one millionth of a joule. The energy
needed to keep a 100-watt bulb lit for an hour is 360,000 joules).
However, the energy costs of moving the data needed to do a single
calculation moving 200 bits of data in and out of memory multiple
times might cost anywhere from 1,000 to 10,000 picojoules.
Message sent using MelbPC WebMail Server
More information about the Link