If Moore's Law hits a sinkhole...

[tw]

Previously posted was that Moore's Law for computers may be about to hit a brick wall. Fundamental to the problem is that transistors have shrunk to about as small as atomic physics permits. On 15 Aug 2005, EE Times wrote this front page story:

Pity the poor MOSFET. Once the star of microelectronics - the ideal blend of elegant simplicity that made the global industry possible - the planar metal-oxide-semiconductor field-effect transistor appears to be approaching the end of its useful life. It leaks where it shouldn't. It's plugged up where it should be open. And repeated attempts to keep it functioning have left the vital parts dangerously thin.

But if the planar MOSFET is in trouble, what comes next? Some experts say there really is no replacement - that new materials, more advanced fabrication techniques, maybe a new substrate will keep the little guy going well into the next decade. Others believe only a move to a radically different, three dimensional structure can keep device scaling on track much beyond the 45 nanometer chip manufacturing mode ...

High-k dielectric materials were suppose to ride to the rescue about now, improving the electric-field strength in the channel and thus reducing leakage for a given dielectric thickness ...

But both high-k and metal gate technologies have proven far from easy to integrate. "This path calls for new, hostile materials to be introduced into the process flow, each with its own fundamental problems.

This materials include Zirconium oxides, Hafnium oxides, or Lanthanum oxides. Not commonly known materials because many are rare as well as little understood.

The upcoming Intel processors will soon be doing 45 nanometer transistors. As the EE Times article points out, simulations for a 32 nanometer transistor result in massive leakage problems. You have already seen how bad the leaking is. Hold you finger on a Pentium processor without a heatsink. No, the Pentium will not burn up as so many claim. But your finger will. Transistors are already so thin that Pentiums can produce more heat than a 100 watt light bulb.

Intel recently announced a fundamental change in design philosophy. No big surprise. Everyone was predicting it. Traditionally, we put all critical computer functions in one chip so that communication between those functions is fastest. Once we send a signal between two ICs, the message gets really slow. However Intel must now separate CPU functions - because transistors are leaking and therefore consuming so much electricity. The trick is to separate CPU functions that need not communicate so fast. That literally means changing the entire CPU architecture. Last time a change this radical was performed, it was called the original Pentium.

Intel has also set new corporate limits on how much electricity a chip can consume. IOW CPU speed is no longer the mantra. Once it was all about doing the same computations only faster. Trying to maximize CPU processing verses minimal electric consumption is a new corporate wide directive. All this to avoid the brick wall.

I am reminded of the last scenes in James Bond's Thunderball where the hydrofoil is desperately swerving to avoid rock outcroppings. Eventually, a crash into rocks as hard as bricks. In the CPU business, are we about to witness the same end? And who will get the girl? Oh. Back then James never got the girl. He was rescued. And then he was replaced by Roger Moore. Curious coincidence.