ADVANTAGES

 

Modern technology is characterized by a continuing search for methods capable of making devices that provide higher performance, smaller size, lower cost, longer independent operation and improved energy efficiency.  The cost of faster devices has been high in power as well as hardware. 

 

Finesse will replace conventional systems by an innovative, scalable fabric of distributed fuel cells commingled with logic.  Each fuel cell will provide power to one logical domain.  A seamless flow of logic will be supported across domain boundaries.  The use of distributed self-biasing logic/power modules will:

 

(i)                  Increase device functional density;

(ii)                Decrease power distribution complexity;

(iii)               Decrease guard bands;

(iv)              Reduce manufacturing steps;

(v)                Decrease power requirements;

(vi)              Improve reliability;

(vii)             Reduce assembly costs. 

 

The design rules for the fuel cells will be expressed in hardware description language (HDL), so that integrated fuel cell/logic modules can be designed, compiled and then produced using standard wafer processing steps.  The economies of commercial mass fabrication in a foundry will then extend to complete devices.  The result will be general computing systems that can operate with higher performance purely from fuels and oxidants. 

 

A problem with current technology is that the power requirement is on a per-cycle basis.  Each clock-cycle takes one current pulse from the power source.  The pulse sizes differ, but the maximum value in a complex CPU can be high.  The distribution of power to support large sets of logical gates has become challenging because the voltages and their tolerances continue to decrease while the worst-case currents continue to increase.  Finesse has developed a completely new paradigm: Small groups of gates have smaller ratios of worst case to average inrush current and are independent of the surrounding logic.

 

This paradigm presents a means of reducing the power distribution problem. It produces the required bias voltage in a distributed manner.  The bias of small sets of gates has small distribution loss.  The power is supplied by distribution of fuel and oxidant.  The fuel source is of high energy density and content.  The bias of such small sets of components produces little noise.  Fuel is depleted mainly as power is drawn rather than through regulators and lossy passive elements so system power is low. The high energy density and low energy use will lead to increased operational lifetimes and/or functionality. 

 

This fuel cell technology contributes a system of integrating a power source on a logic wafer and removes many connections and conditioning functions involved in fitting complex systems into smaller spaces.  The development of ever-smaller nano-scale electronic elements requires new solutions to be identified with respect to routing power while isolating various types of unwanted signaling paths.  These currently involve costly additional process steps.  This distributed power concept will provide a new degree of freedom for designing and compiling more versatile electronic devices.  The ubiquitous presence of microelectronics in society renders a tremendous application potential for the proposed modules.  Within the field of portable microelectronics; laptop computers, mobile communication devices, satellites for reconnaissance or communication, space land rovers, and deferred deployment systems, can all benefit from this technology.  The method with the least constraints inevitably becomes pervasive, just as integrated circuits have supplanted discrete component designs.