To clear confusion, "parallel Hz" has to do with *clock rate* not with *bit rate*. The bits may or may not be parallel. However, the Hz is parallel. Each Hz has its own channel. A billion of these channels result in 1 GHz. Now the bits themselves maybe serial but the Hz isn't. Each channel is limited to 1 Hz.
Parallel Hz -- Confusion
Started by ●August 24, 2004
Reply by ●August 24, 20042004-08-24
Curious wrote:> To clear confusion, "parallel Hz" has to do with *clock rate* not with > *bit rate*. > > The bits may or may not be parallel. However, the Hz is parallel. Each > Hz has its own channel. A billion of these channels result in 1 GHz. > Now the bits themselves maybe serial but the Hz isn't. Each channel is > limited to 1 Hz.So?
Reply by ●August 24, 20042004-08-24
Jerry Avins <jya@ieee.org> wrote in message news:<412bb347$0$21738$61fed72c@news.rcn.com>...> So?Its safer from burn-out than 1 channel with 1 GHz. 1 billion channels each at 1 Hz won't get as hot as 1 channel at 1 GHz. Why not replace "serial Hz" with "parallel Hz"?
Reply by ●August 24, 20042004-08-24
Curious wrote:> Jerry Avins <jya@ieee.org> wrote in message news:<412bb347$0$21738$61fed72c@news.rcn.com>... > > >>So? > > > Its safer from burn-out than 1 channel with 1 GHz. 1 billion channels > each at 1 Hz won't get as hot as 1 channel at 1 GHz. Why not replace > "serial Hz" with "parallel Hz"?One billion channels at one Hz will develop as much heat (simplistic!) as one channel at a GHz. In the same volume, they will reach the same temperature. How many channels can you afford to lose before your gizmo gives wrong answers? Do channels communicate with one another? How? Please describe how you would add two 32-bit numbers and estimate how long it would take. How long does it take on the computer you're using? You look at the big picture (with bad physics), but the devil is in the details. Waving your arms won't do. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●August 25, 20042004-08-25
Curious wrote:> Jerry Avins <jya@ieee.org> wrote in message news:<412bb347$0$21738$61fed72c@news.rcn.com>... >>So? > > > Its safer from burn-out than 1 channel with 1 GHz. 1 billion channels > each at 1 Hz won't get as hot as 1 channel at 1 GHz. Why not replace > "serial Hz" with "parallel Hz"?What makes you think that heat dissipation is proportional to Hz, rather than to bit transitions? There are also issues of skew and physical size. -- Andrew
Reply by ●August 25, 20042004-08-25
Jerry Avins <jya@ieee.org> wrote in message news:<412bf8be$0$21738$61fed72c@news.rcn.com>...> Curious wrote: > > > Jerry Avins <jya@ieee.org> wrote in message news:<412bb347$0$21738$61fed72c@news.rcn.com>... > > > > > >>So? > > > > > > Its safer from burn-out than 1 channel with 1 GHz. 1 billion channels > > each at 1 Hz won't get as hot as 1 channel at 1 GHz. Why not replace > > "serial Hz" with "parallel Hz"? > > One billion channels at one Hz will develop as much heat (simplistic!) > as one channel at a GHz. In the same volume, they will reach the same > temperature. > > How many channels can you afford to lose before your gizmo gives wrong > answers? Do channels communicate with one another? How? Please describe > how you would add two 32-bit numbers and estimate how long it would > take. How long does it take on the computer you're using? > > You look at the big picture (with bad physics), but the devil is in the > details. Waving your arms won't do.Don't forget the ~10 GHz clock needed to syncrionize and phase align all these 1 billion channels. And all the wiring needed to sync up this thing. With some luck, you might end up with a device that can just fit inside the volume taken up by the ENIAC. In such a situation, arm-waving would be of the good. The cooling fans would work over-time to ensure sufficient circulation of air. Rune
Reply by ●August 25, 20042004-08-25
Jerry Avins wrote:> > You look at the big picture (with bad physics), but the devil is in the > details. Waving your arms won't do.I think perhaps Curious should try to apply his technology to the transportation industry. He could design a car with a billion wheels instead of four. With that car he could get to where he was going before he even left where he was coming from. -jim -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =-----
Reply by ●August 25, 20042004-08-25
"Jerry Avins" <jya@ieee.org> wrote in message news:412bf8be$0$21738$61fed72c@news.rcn.com...> Curious wrote: > > > Jerry Avins <jya@ieee.org> wrote in messagenews:<412bb347$0$21738$61fed72c@news.rcn.com>...> > > > > >>So? > > > > > > Its safer from burn-out than 1 channel with 1 GHz. 1 billion channels > > each at 1 Hz won't get as hot as 1 channel at 1 GHz. Why not replace > > "serial Hz" with "parallel Hz"? > > One billion channels at one Hz will develop as much heat (simplistic!) > as one channel at a GHz. In the same volume, they will reach the same > temperature. > > How many channels can you afford to lose before your gizmo gives wrong > answers? Do channels communicate with one another? How? Please describe > how you would add two 32-bit numbers and estimate how long it would > take. How long does it take on the computer you're using? > > You look at the big picture (with bad physics), but the devil is in the > details. Waving your arms won't do.But Julius says "The most rigorous proofs will be shown by vigorous handwaving" :-)) Cheers Bhaskar> > Jerry > -- > Engineering is the art of making what you want from things you can get. > ����������������������������������������������������������������������� >
Reply by ●August 25, 20042004-08-25
"Curious" <curious11112001@yahoo.com> wrote in message news:34a4f456.0408241232.889c34a@posting.google.com...> To clear confusion, "parallel Hz" has to do with *clock rate* not with > *bit rate*. > > The bits may or may not be parallel. However, the Hz is parallel. Each > Hz has its own channel. A billion of these channels result in 1 GHz. > Now the bits themselves maybe serial but the Hz isn't. Each channel is > limited to 1 Hz.Except for the relatively ridiculous numbers you're using (a ratio of 1 Billion only serves to encite skepticism), the method you present is well-known as space-division multiplexing. That is, one physical channel goes to many and, in reverse, many physical channels compress onto one. There are lots of examples: When dialup telephone lines limit out at 56kbps, then using multiple lines at one user site has been applied to increase available bandwidth. There have been modems on the market to achieve such a 2-line capability. Some sort of time division multiplexing is necessary to eventually combine what is received from the two lines. The converse is also interesting: When multiple users are connected physically to individual lines with modest capacity, then time or frequency division multiplexing is used to achieve spatial compression (i.e. to put all the user's data onto one line like a T1). There's no magic or new invention here in moving from physical separation to physical combinations or vice versa. There is simply good engineering to achieve a desired end result using what's available and physically realizable. That is always a matter of making trades and judgements. And, this is the place to put your energy if you're putting forth a new idea. So far, I don't see a new idea - but I do try to keep an open mind. I hope this is helpful in putting the notion into some context. Fred
Reply by ●August 25, 20042004-08-25
jim <"N0sp"@m.sjedging@mwt.net> wrote in message news:<412c6bb5_1@corp.newsgroups.com>...> He could design a car with a billion wheels > instead of four.Actually the correct car analogy would be 1 billion engines each revving at 1 rpm to gain an overall speed of 1 billion rpm. In this car, a lower gear would use more engine revving "channels" to obtain the same mph than a higher gear would. 1 rpm = 1 channel 1 billion rpm = 1 billion channels
