Pushmi-Pullyu <PullmiPushyu@[EMAIL PROTECTED]
> wrote:
> Anyway, this is pretty darn cool...
>
> P
>
> http://tinyurl.com/59xql3
> "From The Sunday Times
> April 6, 2008
> Coming soon: superfast internet
> Jonathan Leake, Science Editor
>
> THE internet could soon be made obsolete. The scientists who pioneered
> it have now built a lightning-fast replacement capable of downloading
> entire feature films within seconds.
>
This article seems to have been written by someone with a poor grasp of
technology, and who mixes several different ideas together
indistinguishably. What CERN is doing is not anything particularly new
(well, the LHC is new, but the computing techniques are not, other than
perhaps in scale). And for the average person, much of what they are
doing is not relevant. I don't do anything on a computer that requires
processing of hundreds of gigabytes per second, and if I did, I wouldn't
get a benefit out of cloud computing unless I had a much higher bandwidth
connection to utilize. I can't download "an entire feature film within
seconds"...not because I lack the computing resources, and not because
the internet doesn't transmit all the data in parallel among many routes,
but because there is a bottleneck where data comes into my
neighborhood/house.
For tasks with very high computation needs, the average person is going
to see the biggest benefit over the next few years not from cloud
computing (sharing the load among computers all over the world) or
internet parallelism (transmitting data concurrently along many paths
simultaneously). It will come from continued evolution in CPU power and a
transition to using graphics processors (GPUs) to do some of the tasks
that are now relegated to CPUs but which GPUs are much better suited to
perform. For example, ripping a DVD or transcoding its format from native
DVD to whatever your ****table media player uses could be done 10-100x
faster if done using a GPU. And that GPU is mostly sitting there unused
right now.
Tasks that require high bandwith are going to steadily improve, but the
barriers to those improvements are the same economic barriers that have
always existed, but is less relevant to CERN and the LHC. If someone
would invent nanites that would build a buried high-speed fiber optic
cable along any desired path, THAT would be cool.
--
"I absolutely promise you that "genius" isn't what most people see me
as." - Ulo Melton
|
