A Study on Nvidia Cuda Architecture and Implementation of Parallel Sat Solver.

Date of Submission

December 2011

Date of Award

Winter 12-12-2012

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science

Department

Advance Computing and Microelectronics Unit (ACMU-Kolkata)

Supervisor

Das, Nabanita (ACMU-Kolkata; ISI)

Abstract (Summary of the Work)

In spite of the enormous progress in the performance of SAT solvers in recent years, still there is strong demand for highly efficient SAT algorithms to solve harder and larger problems. Though there exists huge scope of parallelism, unfortunately, most modern SAT solvers are sequential. Starting with a concise report on CUDA architecture and programming basics, this dissertation presents an implementation of parallel matrix multiplication algorithm on NVIDIA CUDA GPU. Average performance is evaluated executing it on large number of random matrices, in terms of time of completion and speed-up varying the scale of parallelism. Next, this parallel matrix multiplication algorithm is used to solve the Satisfiability problem. In present work, both complete and incomplete SAT solvers have been considered and parallel algorithms are developed and average performance is evaluated executing the algorithms on NVIDIA CUDA GPU for large number of randomly generated Boolean functions.

Comments

ProQuest Collection ID: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:28843254

Control Number

ISI-DISS-2011-290

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/6447

This document is currently not available here.

Share

COinS