Home         Authors   Papers   Year of conference   Themes   Organizations        To MES conference

Use of parallel computing in VLSI computer-aided design  

Authors
 Cheremisinov D.I.
 Cheremisinova L.D.
Date of publication
 2016

Abstract
 The paper is devoted to the problem of enhancement of efficiency of CAD tools for the sake of computational speedup of combinatorial tasks underlying by means of code parallelization. The goal of the paper is twofold: first, propose a framework simplifying the creation of distributed software for the supercomputer family “SKIF” [2], that is Beowulf cluster [5] having master-worker architecture and consisting of a large number of interconnected stand-alone (with individual memory) computational nodes and a master front-end node. The nodes run under Linux and collaborate through master node using MPI [6].
The second goal is to show how the parallel calcula-tions conducted by the supercomputer can be integrated with CAD tools implemented in Windows environment.. As solving logical equation in the form of conjunctive normal form (Boolean satisfiability – SAT) is the most often used as the underlying model for a significant and increasing number of EDA algorithms [1] its parallel im-plementation in the cluster like computing environment was suggested (similar as in [16]).
The methodology and software are suggested which provide data compatibility and interworking between clus-ter computer of Beowulf type and CAD system that is intended for solving laborious tasks of logical design. The suggested grid service is three-layer system consisting of MPI-program for supercomputer, link services and user agent, and it is based on using open protocols and middle-ware.
As an example, the technology and results of fulfilling verification [3, 20] within the environment of program tool of logical design of custom CMOC VLSI [4] was investigated. Verification was based on satisviability problem solving that has been accomplished in remote access by the developed parallel program using high-performance supercomputer.
Keywords
 design automation, parallel computing, cluster computer, grid system, verification, Boolean satisfiability
Library reference
 Cheremisinov D.I., Cheremisinova L.D. Use of parallel computing in VLSI computer-aided design // Problems of Perspective Micro- and Nanoelectronic Systems Development - 2016. Proceedings / edited by A. Stempkovsky, Moscow, IPPM RAS, 2016. Part 1. P. 32-39.
URL of paper
 http://www.mes-conference.ru/data/year2016/pdf/D028.pdf

Copyright © 2009-2024 IPPM RAS. All Rights Reserved.

Design of site: IPPM RAS