2020/08/17 by Sairam Sri Vatsavai, Vatsavai, Sairam Sri, Venkata Sai Praneeth Karempudi +3
Computer Science · Engineering · #Emerging Technologies (cs.ET) #FOS: Computer and information sciences #FOS: Electrical engineering #Neural Networks and Reservoir Computing #Optical Network Technologies #Photonic and Optical Devices #Signal Processing (eess.SP) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2008.07566
openalex publication_date 2020/08/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The performance of on-chip communication in the state-of-the-art multi-core\nprocessors that use the traditional electron-ic NoCs has already become\nseverely energy-constrained. To that end, emerging photonic NoCs (PNoC) are\nseen as a po-tential solution to improve the energy-efficiency (performance per\nwatt) of on-chip communication. However, existing PNoC designs cannot realize\ntheir full potential due to their exces-sive laser power consumption. Prior\nworks that attempt to improve laser power efficiency in PNoCs do not consider\nall key factors that affect the laser power requirement of PNoCs. Therefore,\nthey cannot yield the desired balance between the reduction in laser power,\nachieved performance and energy-efficiency in PNoCs. In this paper, we present\nPROTEUS framework that employs rule-based self-adaptation in PNoCs. Our\napproach not only reduces the laser power consumption, but also minimizes the\naverage packet latency by opportunis-tically increasing the communication data\nrate in PNoCs, and thus, yields the desired balance between the laser power\nre-duction, performance, and energy-efficiency in PNoCs. Our evaluation with\nPARSEC benchmarks shows that our PROTEUS framework can achieve up to 24.5% less\nlaser power consumption, up to 31% less average packet latency, and up to 20%\nless energy-per-bit, compared to another laser power management technique from\nprior work.\n