Work place: Department of ECE, GITAM University, Hyderabad, Telangana- 502329, India
E-mail: manjunath4005@gmail.com
Website:
Research Interests:
Biography
Dr. K. Manjunathachari Professor & Head of ECE, Department of ECE School of Technology, GITAM University, Hyderabad. B.E (ECE), Gulburga University, Gulburga, completed in 1991. M.Tech (DSCE), JNT University, Hyderabad in 1996. Ph D: JNT University, Kakinada, completed in 2010. On Image processing. Experience: 26years.
By T. Pullaiah K. Manjunathachari B. L. Malleswari
DOI: https://doi.org/10.5815/ijcnis.2025.02.02, Pub. Date: 8 Apr. 2025
Due to the maximal transistor count, Multi-Processor System-on-Chip (MPSoC) delivers more performance than uniprocessor systems. Network on Chip (NoC) in MPSoC provides scalable connectivity compared to traditional bus-based interconnects. Still, NoC designs significantly impact MPSoC design as it increases power consumption and network latency. A solution to this problem is packet compression which minimizes the data redundancy within NoC packets and reduces the overall power consumption of the whole network by minimizing a data packet size. Latency and overhead of compressor and decompressor require more memory access time, even though the packet compression is good for the improved performance of NoC. So, this problem demands a simple and lightweight compression method like delta compression. Consequently, this research proposes a new delta-difference Hybrid Tree coding (∆DHT-Zip) to de/compress the data packet in the NoC framework. In this compression approach, the Delta encoding, Huffman encoding and DNA tree (deoxyribonucleic acid) coding are hybridized to perform the data packet de/compression approach. Moreover, a time series approach named Run Length Encoding (RLE) is used to compress the metadata obtained from both the encoding and decoding processes. This research produces decreased packet loss and significant power savings by using the proposed ∆DHT-Zip method. The simulation results show that the proposed ∆DHT-Zip algorithm minimizes packet latency and outperforms existing data compression approaches with a mean Compression Ratio (CR) of 1.2%, which is 79.06% greater than the existing Flitzip algorithm.
[...] Read more.Subscribe to receive issue release notifications and newsletters from MECS Press journals