Introduction to parallel computing: design and analysis of algorithms
Introduction to parallel computing: design and analysis of algorithms
Communications of the ACM
Combining static and dynamic defect-tolerance techniques for nanoscale memory systems
Proceedings of the 2007 IEEE/ACM international conference on Computer-aided design
| Hi-index | 0.08 |
We present a parallel version of Viterbi's decoding procedure, for which we are able to demonstrate that the resultant task graph has restricted complexity in that the number of communications to or from any processor cannot exceed four for BCH codes. The resulting algorithm works in lock step making it suitable for implementation on a systolic processor array, which we have implemented on a field programmable gate array and have demonstrated the perfect scaling of the algorithm for two exemplar BCH codes. The parallelisation strategy is applicable to all cyclic codes and convolution codes. We also present a novel method for generating the state transition diagrams for these codes.