On 4/1/2014 3:27 PM, Poquito Loco wrote:
On Tue, 01 Apr 2014 13:37:57 -0400, wrote:

On Tue, 01 Apr 2014 13:22:24 -0400, "Mr. Luddite"
wrote:

On 4/1/2014 12:45 PM, wrote:
On Tue, 01 Apr 2014 12:00:08 -0400, "Mr. Luddite"
wrote:

On 4/1/2014 11:38 AM, wrote:


You are still hitting the wall.
Regular chips are about tapped out.
We are rapidly approaching the point that we will be super cooling
processors to get quantum effects.
There is only so much you can do to shorten the data path.
They are just making them wider. (multiple processors, wider buses)


Ummm ... I don't claim to be a semiconductor manufacturing expert nor
have a lot of experience in wafer fab but there are companies investing
a lot of research money into the optical replacement of copper tracing
of single, double and multi-level boards. The focus ( no pun intended)
is on reducing size and complexity. Not sure what gains in overall
processing speeds are achieved although claims are made that it will.

These are tiny, pin head sized laser diodes. The cool thing is that the
light paths can intersect others with no interference or "shorts".

I have read about it in the trade rags. It still seems to have the
intent of making shorter and marginality faster data paths.
When you are splitting hairs on the speed of light vs electrons on
copper, in a chunk of real estate the size of your thumbnail, there is
not much more speed to be had.
Now when they get this quantum computing thing going, they are off to
the races again. I doubt you will have that on your desk anytime soon.


The available bandwidth of an optical system is orders of magnitude
greater than that of copper conductors. Hence, more data can be moved
faster simultaneously.


AKA a wider data path. ;-)

Well...that *was* a bit funny!



It has absolutely nothing to do with the physical size of the "path",
copper or laser beam.

On Tue, 01 Apr 2014 15:47:40 -0400, "Mr. Luddite" wrote:

On 4/1/2014 3:27 PM, Poquito Loco wrote:
On Tue, 01 Apr 2014 13:37:57 -0400, wrote:

On Tue, 01 Apr 2014 13:22:24 -0400, "Mr. Luddite"
wrote:

On 4/1/2014 12:45 PM, wrote:
On Tue, 01 Apr 2014 12:00:08 -0400, "Mr. Luddite"
wrote:

On 4/1/2014 11:38 AM, wrote:


You are still hitting the wall.
Regular chips are about tapped out.
We are rapidly approaching the point that we will be super cooling
processors to get quantum effects.
There is only so much you can do to shorten the data path.
They are just making them wider. (multiple processors, wider buses)


Ummm ... I don't claim to be a semiconductor manufacturing expert nor
have a lot of experience in wafer fab but there are companies investing
a lot of research money into the optical replacement of copper tracing
of single, double and multi-level boards. The focus ( no pun intended)
is on reducing size and complexity. Not sure what gains in overall
processing speeds are achieved although claims are made that it will.

These are tiny, pin head sized laser diodes. The cool thing is that the
light paths can intersect others with no interference or "shorts".

I have read about it in the trade rags. It still seems to have the
intent of making shorter and marginality faster data paths.
When you are splitting hairs on the speed of light vs electrons on
copper, in a chunk of real estate the size of your thumbnail, there is
not much more speed to be had.
Now when they get this quantum computing thing going, they are off to
the races again. I doubt you will have that on your desk anytime soon.


The available bandwidth of an optical system is orders of magnitude
greater than that of copper conductors. Hence, more data can be moved
faster simultaneously.


AKA a wider data path. ;-)

Well...that *was* a bit funny!



It has absolutely nothing to do with the physical size of the "path",
copper or laser beam.

I believe you, but it was just humorous. You referred to 'available bandwidth' being greater, and
Greg talked about a 'wider path'. Well, to a rank amateur like me, 'greater bandwidth' and 'wider
path' sound pretty similar!

Believe me, I wasn't trying to impugn anything you said. When y'all get into the technical stuff, I
keep well out of it.

On Tue, 01 Apr 2014 16:02:33 -0400, Poquito Loco
wrote:

I believe you, but it was just humorous. You referred to 'available bandwidth' being greater, and
Greg talked about a 'wider path'. Well, to a rank amateur like me, 'greater bandwidth' and 'wider
path' sound pretty similar!

===

There are basically two ways to achieve greater bandwidth. One is to
send data at higher speed in a single stream. That works but it is
presently running up against the speed of light, as well as density
and cooling issues. The second way is to break up the data into
multiple parallel streams, i.e., "a wider path", sort of like
converting a two lane road into a 3 or 4 lane road so it can handle
more cars.

On Tuesday, April 1, 2014 5:53:31 PM UTC-4, Wayne. B wrote:
On Tue, 01 Apr 2014 16:02:33 -0400, Poquito Loco

wrote:



I believe you, but it was just humorous. You referred to 'available bandwidth' being greater, and

Greg talked about a 'wider path'. Well, to a rank amateur like me, 'greater bandwidth' and 'wider

path' sound pretty similar!



===



There are basically two ways to achieve greater bandwidth. One is to

send data at higher speed in a single stream. That works but it is

presently running up against the speed of light, as well as density

and cooling issues. The second way is to break up the data into

multiple parallel streams, i.e., "a wider path", sort of like

converting a two lane road into a 3 or 4 lane road so it can handle

more cars.

Early in my career, I laid out PC boards. Did it on a light table with tape, pre-cut pads and an Exacto knife. A lot was audio, with some microproccessor and logic control. I'm glad I moved on before things got so fast that you had to worry about electrons running off the end of right angle copper trace runs. The audio was interesting... lots of games played to reduce crosstalk between paths and to improve S/N. That was fun.

On 4/1/2014 5:53 PM, Wayne.B wrote:
On Tue, 01 Apr 2014 16:02:33 -0400, Poquito Loco
wrote:

I believe you, but it was just humorous. You referred to 'available bandwidth' being greater, and
Greg talked about a 'wider path'. Well, to a rank amateur like me, 'greater bandwidth' and 'wider
path' sound pretty similar!

===

There are basically two ways to achieve greater bandwidth. One is to
send data at higher speed in a single stream. That works but it is
presently running up against the speed of light, as well as density
and cooling issues. The second way is to break up the data into
multiple parallel streams, i.e., "a wider path", sort of like
converting a two lane road into a 3 or 4 lane road so it can handle
more cars.


In the old days we called it "multiplexing" of which there are many
forms or types. The same "road" is used but is shared in terms of the
time it is used.

In RF communications systems capacity is frequency dependent. The
higher the frequency, the more data can be transmitted over the same
"road". Optical systems are orders of magnitude higher in "frequency"
and are expressed in wavelengths and the capacity is increased
correspondingly. Multiple "connections" to a processor that required
several physical roads can be combined into one also.

The other benefit (as previously mentioned) is the ability for optical
paths to cross and intersect, unlike physical copper tracing.

On Wed, 02 Apr 2014 00:22:39 -0400, wrote:

On Tue, 01 Apr 2014 22:24:26 -0400, "Mr. Luddite"
wrote:

On 4/1/2014 5:53 PM, Wayne.B wrote:
On Tue, 01 Apr 2014 16:02:33 -0400, Poquito Loco
wrote:

I believe you, but it was just humorous. You referred to 'available bandwidth' being greater, and
Greg talked about a 'wider path'. Well, to a rank amateur like me, 'greater bandwidth' and 'wider
path' sound pretty similar!

===

There are basically two ways to achieve greater bandwidth. One is to
send data at higher speed in a single stream. That works but it is
presently running up against the speed of light, as well as density
and cooling issues. The second way is to break up the data into
multiple parallel streams, i.e., "a wider path", sort of like
converting a two lane road into a 3 or 4 lane road so it can handle
more cars.


In the old days we called it "multiplexing" of which there are many
forms or types. The same "road" is used but is shared in terms of the
time it is used.

In RF communications systems capacity is frequency dependent. The
higher the frequency, the more data can be transmitted over the same
"road". Optical systems are orders of magnitude higher in "frequency"
and are expressed in wavelengths and the capacity is increased
correspondingly. Multiple "connections" to a processor that required
several physical roads can be combined into one also.

The other benefit (as previously mentioned) is the ability for optical
paths to cross and intersect, unlike physical copper tracing.

You still have to keep those pipes full and the only way they can do
it is with a bunch of processors because they have hit the wall on how
fast they can go,

That is why we don't see faster processor speeds advertised anymore.
They just talk about how many "cores" they have.

===

Yes, and now we need more software apps that are capable of using
those cores effectively. Unfortunately XP does not do a good job
supporting multi-cores either. It is currently, and most probably
always, limited to two cores if my memory is correct.

On 4/2/14, 10:56 AM, wrote:
On Wed, 02 Apr 2014 00:48:04 -0400, Wayne.B
wrote:

On Wed, 02 Apr 2014 00:22:39 -0400, wrote:

That is why we don't see faster processor speeds advertised anymore.
They just talk about how many "cores" they have.

===

Yes, and now we need more software apps that are capable of using
those cores effectively. Unfortunately XP does not do a good job
supporting multi-cores either. It is currently, and most probably
always, limited to two cores if my memory is correct.

If I get a quad core machine I suppose I will need newer software, I
understand that but I do not need that extra speed for anything I do.
I doubt most people do either but they just want the next new thing
for some reason.
Harry's only excuse is he saves a few seconds ripping DVDs he will
never watch to the humongous file server he bought, just to fill it up
I suppose.



I offered up one example of an app I use, and you build your negative
universe off of that one example? Some science guy you are.

It's not a few seconds, by the way. My desktop computer transcodes DVDs
in about half the time of the Windows computer I used to use. That a
savings of at least 15 minutes on each transcode. Part of it is
software, part of it is hardware, part of it is the Apple OS.

And, in fact, I do watch many of my old favorite movies a couple of
times a year. They only take up a small portion of the space available
on our "humongous" file server. Of course, we don't have to have a half
dozen antique computers flopped around the house, eh?

My wife prefers to run her Mathematica9 projects on my iMac instead of
on her Win 7 i5 machine, which is only about 18 months old. Calcs and
procedures run a hell of a lot faster on the Mac.

On 4/2/14, 11:41 AM, wrote:
On Wed, 02 Apr 2014 11:15:07 -0400, F*O*A*D wrote:

On 4/2/14, 10:56 AM, wrote:
On Wed, 02 Apr 2014 00:48:04 -0400, Wayne.B
wrote:

On Wed, 02 Apr 2014 00:22:39 -0400, wrote:

That is why we don't see faster processor speeds advertised anymore.
They just talk about how many "cores" they have.

===

Yes, and now we need more software apps that are capable of using
those cores effectively. Unfortunately XP does not do a good job
supporting multi-cores either. It is currently, and most probably
always, limited to two cores if my memory is correct.

If I get a quad core machine I suppose I will need newer software, I
understand that but I do not need that extra speed for anything I do.
I doubt most people do either but they just want the next new thing
for some reason.
Harry's only excuse is he saves a few seconds ripping DVDs he will
never watch to the humongous file server he bought, just to fill it up
I suppose.



I offered up one example of an app I use, and you build your negative
universe off of that one example? Some science guy you are.

It's not a few seconds, by the way. My desktop computer transcodes DVDs
in about half the time of the Windows computer I used to use. That a
savings of at least 15 minutes on each transcode. Part of it is
software, part of it is hardware, part of it is the Apple OS.

so you rip some DVDs that is what I said

Among other things.


And, in fact, I do watch many of my old favorite movies a couple of
times a year. They only take up a small portion of the space available
on our "humongous" file server. Of course, we don't have to have a half
dozen antique computers flopped around the house, eh?

How do these movies get from the file server to your TV?


Through the magic of wi-fi. Modern computers, servers, TV sets, DVD
players, and boxtop devices have wired/wireless router hookup
capabilities, and that means they can communicate with each other.
Perhaps you can trade in that 300 bps Hayes...


\
My wife prefers to run her Mathematica9 projects on my iMac instead of
on her Win 7 i5 machine, which is only about 18 months old. Calcs and
procedures run a hell of a lot faster on the Mac.


Sure

Oh, they do. My wife works a lot with math, stats, et cetera. She took a
*lot* of upper level math and stat courses to get her M.S. and Ph.D, and
use them to help along various research projects in which she is involved.

On 4/2/14, 12:22 AM, wrote:
On Tue, 01 Apr 2014 22:24:26 -0400, "Mr. Luddite"
wrote:

On 4/1/2014 5:53 PM, Wayne.B wrote:
On Tue, 01 Apr 2014 16:02:33 -0400, Poquito Loco
wrote:

I believe you, but it was just humorous. You referred to 'available bandwidth' being greater, and
Greg talked about a 'wider path'. Well, to a rank amateur like me, 'greater bandwidth' and 'wider
path' sound pretty similar!

===

There are basically two ways to achieve greater bandwidth. One is to
send data at higher speed in a single stream. That works but it is
presently running up against the speed of light, as well as density
and cooling issues. The second way is to break up the data into
multiple parallel streams, i.e., "a wider path", sort of like
converting a two lane road into a 3 or 4 lane road so it can handle
more cars.


In the old days we called it "multiplexing" of which there are many
forms or types. The same "road" is used but is shared in terms of the
time it is used.

In RF communications systems capacity is frequency dependent. The
higher the frequency, the more data can be transmitted over the same
"road". Optical systems are orders of magnitude higher in "frequency"
and are expressed in wavelengths and the capacity is increased
correspondingly. Multiple "connections" to a processor that required
several physical roads can be combined into one also.

The other benefit (as previously mentioned) is the ability for optical
paths to cross and intersect, unlike physical copper tracing.

You still have to keep those pipes full and the only way they can do
it is with a bunch of processors because they have hit the wall on how
fast they can go,

That is why we don't see faster processor speeds advertised anymore.
They just talk about how many "cores" they have.



I see computers advertised all the time with spec's showing the speed of
the processor(s).

Under "specifications" on this page, Apple "specifies" what processor is
being used and what its clock speed is -

http://www.apple.com/mac/compare/

Those speeds are higher than in the previous generation of Mac computers.

When you click for more details on specific models, you get this sort of
page:

http://store.apple.com/us/buy-mac/im.../A&step=config

If you click on an HP page, the specs show the processor speeds:

http://tinyurl.com/ls5noup