Review #1

PROPOSAL NO.: 0340256
INSTITUTION: pragc
NSF PROGRAM: SMALL BUSINESS PHASE I
PRINCIPAL INVESTIGATOR: Meyer, Steven J
TITLE: SBIR Phase I: SPICE Circuit Simulator Direct Solution Improvement
RATING:Poor

REVIEW:
What is the intellectual merit of the proposed activity?

The general thrust of this proposal is to improve (upgrade capabilities) an older, well established electrical circuit solving program (SPICE), currently in widespread use.  The need for larger scale, accurate computation of various systems is apparent in many areas (such as electrical circuit development, discussed here) is widely recognized. In some cases older commercial software is well tested and popular but slow and cumbersome when attacking larger scale problems.   It is suggested in this proposal that previously undiscovered and innovative techniques for numerical solutions might be found by examining the work in fundamental physics (in some cases predating modern computational efforts).   Thus, it would appear that the proposal is concerned with developing entirely new algorithms for application.   It is mention several times that SPICE (as currently constituted) is accurate and stable.  While this may be so, a significant shift in the underlying computational methods (to some method yet to be discovered) may well introduce new and unanticipated numerical problems.   The narrative seems to imply that stability should be preserved by some inherent characteristic of SPICE, a leap that is hard to accept.  Given that no specific algorithm has been found or proposed for this application, the total effort proposed here may be inconsistent with the time frame available.  



What are the broader impacts of the proposed activity?

Certainly an updated, efficient and expanded version of SPICE would have considerable commercial appeal.   It is also likely that if improvements in this type of code could be made, then spill-over to parallel applications might well take place.  Aside from the need to attack large scale problems, at least part of the improvements in any code of this type center on the requirements to model complex elements that are not easily represented by tractable equations.  Any generalization to broader markets will need to address this issue.

Summary Statement

Funding for this proposal at Phase I level cannot be recommended, primarily because of the lack of focus on a work plan needed to move a concept towards commercialization.   What is proposed is research intended to find new methods.