Moore's Law has decades left, Intel CTO predicts

Moore's Law will keep going strong for decades, Intel CTO Justin Rattner predicts.

Moore's Law, in force for more than 40 years, says that the number of transistors that can be placed on an integrated circuit will double every 18 to 24 months. Predictions of the demise of Moore's Law are routinely heard in the IT world, and some organizations are trying to find a replacement for silicon chip technology. But Rattner says that silicon has plenty of life left and said there is no end in sight for Moore's Law.

"If Moore's Law is simply a measure of the increase in the number of electronic devices per chip, then Moore's Law has much more time to go, probably decades," Rattner said in an interview with Network World.

My understanding is that they were at the edge of moore’s law. They knew this because the “yield” of CPUs that passed the performance check was so low that it was almost not worth making them anymore.

Then along comes this idea of multiple “cores”. If I get it right, the purpose of the multiple cores was to increase the yield of functional CPUs. You see, if you have a quad core chip, and one or two of the cores don’t meet performance specs, you simply disable the bad cores and then sell it as a “two core” CPU. You don’t get as much money selling it as a double core, but it’s way better than zip.

Essentially, they hit the wall already. Moore’s law has already broke down. Multi core chips is a neat little trick to improve beyond the death of moore’s law.

That’s my understanding anyway.

They have to work on something that will generate income in the meantime. What I mean is, they can’t spend 60 years replicating a cockroach brain with transistors. They have to go in baby steps and at each step of the way they must figure out a way to package the incremental progress into a useful product for sale to the public at a profit.

So I think their only way forward is to try to FOOL us. They will continue to make software that APPROXIMATES intelligence, or replicates it in some very narrow but useful(or at least entertaining) fashion. Maybe some day, they will become so good at fooling us, that they will accidentally discover how to create true intelligence.

I expect the first forms of success will come from more sophisticated phone answering software. Once it’s perfected beyond a certain point, they could start replacing people in all kinds of jobs. fast food drive thrus. Checkout lanes at all kinds of retail stores. gas stations. restaurants. etc.

That’s exactly backwards from the way mother nature appears to have done it. Human thought and human language would seem to be the pinnacle of mother natures(or god’s) creation. While vision(or at least light perception) and navigation would seem to be a much earlier creation.

The main problem in the megahertz race was heat. At one point Intel was trying to dissipate 115W from an area barely over one square centimeter. That's a serious heat concentration.

In addition, the speed meant they needed very long pipelines. Sometimes the electrical signals couldn't propagate across the chip fast enough and instructions would be stuck in essentially a wait stage of the pipeline.

IOW, this idea ran very hot, and wasn't very efficient. Remember, the first Pentium 4 with this architecture did less work than a Pentium 3 of the same clock speed.

If I get it right, the purpose of the multiple cores was to increase the yield of functional CPUs.

That showed up on AMD's three-core chip. This calculation is also the reason the PS3 has only seven of the eight processing units enabled. They disable one even if all eight are good.

But usually the yield quality determines clock speed. Lower quality produces lower clock speed. This was also a reason for the switch, as Intel couldn't get good yields on the high-speed processors, resulting in a glut of many lower speed ones for very few capable of the highest speeds as they were pushing the technology.

. Moore’s law has already broke down.

Moore's law is often misunderstood. Moore said "The complexity for minimum component costs has increased at a rate of roughly a factor of two per year." He predicted this to last 10 years, after which he changed it to every two years (not 18 months).

Seeing the actual definition, it's clear that Moore's Law is alive and well. Two cores being put in a chip for the same cost as a single core two years ago perfectly represents Moore's Law.

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.