Folding @ Home

Fyron · #1 November 8th, 2003, 10:09 PM

Actually... you need variable resistances. You need high resitance for some parts, low for others, but mostly semi-conductors that you can change their resistance. Otherwise, you can't build the necessary circuits.

Newer CPUs do not have less resistance or anything, they just make the parts smaller, so they can cram more on to the CPU board. The problem is, the smaller you make it, more of a problem the electric fields generated by the parts get.

Fiber optics might be one area where they could drastically change CPU architecture. Or those organic computers some people are trying to make...

[ November 08, 2003, 20:10: Message edited by: Imperator Fyron ]

narf poit chez BOOM · #2 November 8th, 2003, 10:12 PM

organic computers? somehow, that doesn't sound good.

deccan · #3 November 9th, 2003, 12:08 AM

Quote:

Originally posted by Imperator Fyron:
Actually... the rate of increase will slow down tremendously.

Apparently the industry consensus is that Moore's law will hold true for at least another decade.

Less is Moore

I like this quote:

Quote:

For one thing, the improvement of chips now faces serious technical challenges. The more densely transistors are packed, the hotter they get. Intel's chips will soon reach the energy density (meaning watts per square centimetre) of a nuclear reactor. Most people do not want a small nuclear reactor on their laps: it will not cause computers to melt down, but laptops have already inflicted nasty burns.

Thermodyne · #4 November 9th, 2003, 01:24 AM

Moore�s Law is already in some jeopardy. Intel has run into leakage problems with the new P4 chip �Prescott�. The transistors are so numerous and packed so tightly in the stretched silicon that they leak more current than the chip actually uses. Over 110 watts for the first production yield. These have been written off and the design team has gone back to work on reducing the leakage. An ominous sign is that Intel has begun to talk publicly about SOI metal hybrid chips for the desktop. These would solve the 90nano leakage problems, but add a very high premium to the price of a CPU. In the past, this technology was seen as only being economical for MilTech applications. For Moore�s Law to continue there will need to be a move away from silicon. Carbon seems to be the front runner at this time.

narf poit chez BOOM · #5 November 9th, 2003, 01:36 AM

Quote:

Intel's chips will soon reach the energy density (meaning watts per square centimetre) of a nuclear reactor.

Thermodyne · #6 November 10th, 2003, 03:55 PM

Nothing like a nice hot Laptop to warm your hands on a cold fall morning

rdouglass · #7 November 10th, 2003, 09:42 PM

Thermo,

Currently giving spare cycles to SETI@home project. I have almost 25K WU's! (Used to see you there - you zoomed by my stats then dissapeared about a year ago from submitting new WU's).

Can you tell me why I should switch? Not looking to start anything - just looking to see if this one is better.

In fact, I think that any machine that has any idle time at all (while it's powered on) should run some kind of "public service" process. It doesn't cost the CPU owner anything (in the vast majority of cases) and helps out good causes...