Scheduling in multi-Wavelength Ring-based Optical Networks-on-Chips
The unprecedented use of electronic gadgets, cloud computing, and web services has pushed the computing systems to its limit in handling huge amount of complex data. Several researches have been conducted to solve this problem and one of these is through the parallelization of cores. The processor a...
Saved in:
Main Author: | |
---|---|
Format: | text |
Published: |
Archīum Ateneo
2020
|
Subjects: | |
Online Access: | https://archium.ateneo.edu/theses-dissertations/412 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Ateneo De Manila University |
Summary: | The unprecedented use of electronic gadgets, cloud computing, and web services has pushed the computing systems to its limit in handling huge amount of complex data. Several researches have been conducted to solve this problem and one of these is through the parallelization of cores. The processor architecture has evolved from single to multiple processing cores working in parallel onto a single die. This paradigm shift has required an efficient interconnection network between cores and processors. Scheduling of packet transmission in synchronous networks-on-chip (NoC) is necessary for obtaining high throughput, low latency, and good fairness, while avoiding packet collisions. Efficient algorithms are present for rearrangeably non-blocking NoC. However, when realized with integrated optical devices, NoC are typically arranged in topologies that are blocking if single wavelength is used. This paper presents a scheduler for an integrated optical NoC based on a ring topology and realized with multiple resonating microring (MMR). Scheduling in MMR architecture consists of the conventional matching sub-problem and the wavelength assignment sub-problem. This paper presents a comprehensive comparison between the two-step scheduler based on the more complex maximum weighted matching algorithm and with an optimal scheduler carried out on both fixed and tunable transmitters. Simulation results indicate that the two-step scheduler performs similarly to the optimal scheduler for most of the scenarios. iPWM has a performance comparable to the two-step schedulers. |
---|