MICROELECTRONIC SYSTEMS NEWS

FILENUMBER: 9808 BEGIN_KEYWORDS CAD EXPERIMENTS END_KEYWORDS DATE: March 1998 TITLE: CAD EXPERIMENTS

=================================================================

TITLE: CAD EXPERIMENTS

=================================================================

If you work in the area of CAD of microelectronic systems or heuristics to
solve NP-hard problems, some items in this message may be of interest. The
purpose is to

(1) alert you to new postings on our web-server, in particular under
EXPERIMENTS .
(2) invite you to participate in the new generation of design and execution of peer-reviewed collaborative experiments, using workflows on the Web. With your participation, we hope to present some of this effort as -- part of a session during International Workshop on Logic Synthesis See: IWLS'98
-- part of the Vela Project demo in the University Booth during DAC'98.
-- part of a joint-session proposal to ICCAD'98 .
The recent first-generation capability to synthesize a large number of equivalence class circuit mutants from a copy of a known design, maintaining either wiring signature invariance or entropy signature invariance, motivated us to conduct a series of statistical experiments with algorithms in layout, logic, and test synthesis. These experiments are beginning to address tests of hypothesis such as "Is the improvement due to the choice of the algorithm or due to chance?" Wide participation and a number of peer-reviewed experiments are required to resolve such issues. The first such discussions may take place at IWLS'98. In particular, thousands of experiments with hundreds of current circuit mutants, posted under EXPERIMENTS.
already raise such issues in logic minimization and variable ordering of ROBDDs. For example, some but not all of the mutants of well-known and relatively simple circuits such as C1355, appear as a serious challenge for ROBDD variable ordering algorithms but not for atpg algorithms, etc. Dr. Franc Brglez Collaborative Benchmarking Laboratory (CBL) Dept. of Computer Science NC State University, Box 7550 Raleigh, NC 27695-7550, USA phone: 919-515-9675 e-mail: brglez@cbl.ncsu.edu fax: 919-513-1895 WWW: Brglez anonymous ftp: ftp.cbl.ncsu.edu autoreply: info@cbl.ncsu.edu =================================================================

Return to MSN Home Page

dbouldin@utk.edu