Korean Job Discussion Forums

[Plutocracy]

Moore's Law (n.) The observation made in 1965 by Gordon Moore, co-founder of Intel, that the number of transistors per square inch on integrated circuits had doubled every year since the integrated circuit was invented. Moore predicted that this trend would continue for the foreseeable future. In subsequent years, the pace slowed down a bit, but data density has doubled approximately every 18 months, and this is the current definition of Moore's Law, which Moore himself has blessed. Most experts, including Moore himself, expect Moore's Law to hold for at least another two decades.

This law has been applied to other areas of technology with success. Assuming that technology in general (especially ICT) will exponentially increase for at least a while yet, what implications do people think this will have on global society in general? Will it aid peace? War? Hurt or harm the environment? At what level are we technologically determined?

Discuss.

[Demophobe]

First, all participants must read this paper.

Then, upon completion, all will discover that it's not really relevant to much outside the computing/technology world (save a few erroneous applications), and even in that, is more about economics than silicon.

Discuss? No, I don't think so. Not here anyways....done it do death in the Ars forums.

I think a more reasonable and interesting discussion would arise from what may stand in the way of Moore's law. Substrates, Ultra-violet lithography, heat dissipation, battery development....some interesting obstacles/compliments; more interesting to me that the (mis)application of an economic observation gome amok.

[Plutocracy]

[Demophobe]

Well, war is always what pushes tech and science forward. War, or the threat of it. Will it continue? Yes. Tech has been going great guns for along, long time. When I say "tech", I don't mean Intel and AMD....I am talking about the whole thing. Ptolomy to Newton....back to the geeks of history.

Is science married to technology? At most stages, yes.

And since the corporations are married to war, as war is what the US economy is arguably built on, and seeing as the US is on the tip of the tech spear (in many ways, but mainly in the consumer, innovation side, leaving the dirty business of silicon and fab to Asia), they will, for the forseeable future, keep this robot walking at a furious pace.

Is that good? Well....no. Not at all. Humans trying to find better ways to kill each other or stop the other dude from killing you is not waht anyone would call a good reason to advance anything. However, many argue that it's our nature to fight, blah, blah.

Sticking to tech.

Storage substrates of late: glass, water, air, plastic. Seems they will put data anywhere.

Lithography: Yes, "extreme" UV is coming...we have no choice, to stay in accordance with Moore's "law". Transistors can only get so small before we start running into problems: heat dissipation, material breakdown under heat...

So, substrates and tooling methods will change soon. Silicon is a real mess to produce, and the cost/return ratios are shrinking, as chip complexity increases and yeilds drop. This is where Moore's "law" breaks down, seeing as it's really an economic idea, rather than a law. As Moore explains, die sizes are already beginning to shrink slower and slower. Right now, we are having a mass of trouble with the 90nm fabbing. As we get smaller, we are seeing limitations in many areas including substrate, heat (from transistor density) and lithography.

We need new "cutting" methods and new stuff to cut. Low-k, strained...we will do anything to try to maximize our returns on silicon. Is it working? Well...graphics. Look at them. They are on the forefront...moreso that CPU chip makers. They are already way past the transistor counts of the CPU makers, and always the first to move to smaller die sizes, new fabbing methods and new types of silicon.

ATI have very poor yields on the X800 chip. It's beginning to break down here. So low in fact, that they have withdrawn the chip from the market and offered cusomers who have already been waiting for months, a "pre-buying" option on their newer chip, which is as yet, unreleased. There is no secret here that it's the fab that is the holdup.

Then, Nvidia found out you need a leaf-blower to cool these things with the 5800 series. It's another breakdown. Though they did 'solve' the problem with a newer design that occupies a large chunk of case real estate and power requrements that make some people like me wonder what kind of 'solution' was that?

Now, with the advent of the "cell" processor, we are seeking to kind of circumvent, or at least offshoot, Moore's law by a method of expansion. Instead of trying to continually add functionality to a single-die design, they moved it to other, off-a-single-die areas of the "cpu". The cell, is in fact, a collection of smaller CPUs, all piggybacking and working in tandem. While this isn't "new, (AMD and Intel have been working on multi-core CPUs for a while, with the first ones coming very soon), the idea of what each of these little areas will be doing is quite unique.

The Cell will certainly solve some of the problems, but may well offer new ones. I don't know enough about the Cell architecture right now to go into it in detail, but I suspect that overall size and power may still be issues. Heat will be remedied, as will the physical barrier of speed. The first Cell is running at 4GHz already.

What does the Cell mean to Moore's law? Well, it kind of breaks it and kind of compliments it. If Moore was indeed talking about pure economics, then the Cell will be in accordance, albeit in a different, multi-die way. If Morre's law is applied to transistor density (and thus speed) on a single die, then the cell is almost flying in it's face. It isn't a single-die solution at all, and thus density problems are alleviated, as are heat and functionality add-on headroom.

The whole heat/power issue is kind of funny. We make CPUs (all silicon generally) smaller to consume less power. Then, we have greatly increased heat production, so we need a cooling solution, which may consume more power by itself that the pre-shrink dies may have.

Enough for now already....my chicken-peck typing style is wearisome.

[Demophobe]

plutocracy...where are your deep replies? By "discuss", did you mean a soloiloquy?

Ah well....good at asking questions, you are.

It's too bad....would have made for an interesting topic.

[Plutocracy]