Boosting the Performance of Interconnection Networks by Selective Data Compression

This study presents a selective data-compression interconnection network to boost its performance. Data compression virtually increases the effective bandwidth. One drawback of data is long latency perform (de-)compression operation at compute node. In terms communication latency, we explore trade-off between overhead and reduced injection by shortening packet length algorithms. As result, present selectively apply technique packet. We packets it also taken when congestion detected source Through cycle-accurate simulation, method using above algorithms improves up 39% throughput with moderate increase in short packets.