| DC 欄位 |
值 |
語言 |
| DC.contributor | 通訊工程學系 | zh_TW |
| DC.creator | 李翊君 | zh_TW |
| DC.creator | Yi-Jun Li | en_US |
| dc.date.accessioned | 2016-7-29T07:39:07Z | |
| dc.date.available | 2016-7-29T07:39:07Z | |
| dc.date.issued | 2016 | |
| dc.identifier.uri | http://ir.lib.ncu.edu.tw:88/thesis/view_etd.asp?URN=103523044 | |
| dc.contributor.department | 通訊工程學系 | zh_TW |
| DC.description | 國立中央大學 | zh_TW |
| DC.description | National Central University | en_US |
| dc.description.abstract | 隨著科技的進步，根據摩爾定律所述，每18個月處理器的效能即為提高一倍。然而，受到物理的限制，在過去經由時脈提升來增加效能的方法已經到達瓶頸。因此，處理器開始由單核心架構轉為往多核心發展。隨著處理器核心數量的增加，傳統的匯流排架構成為多核心處理器中設計與效能的瓶頸。為了解決傳統單晶片架構的問題，將電腦網路中封包交換技術傳遞資料的方式帶入晶片網路架構的設計，而好處是使整體晶片的設計可以更具彈性。與此同時，伴隨無線射頻技術的進步，使得晶片內傳輸透過無線的方式變得成為可能，為下一代的晶片提供高速率、低延遲與低功耗的傳輸方式。隨著光學的進步，也開始於在晶片內加入了波導管的設計，使得晶片間通訊透過光學的方式來達到高速率的傳輸。
在本篇論文中，我們著重在研究二維與三維無線晶片網路架構中各處理器之多核心結構的無線多通道路由演算法，並延伸導入光學波導管的設計優化晶片效能。在基於分群（Group-Based）的方式，研究內部路由協定的設計和通道的分配使用，此方式讓晶片內部的核心進行封包資料的交換，目的在於透過分群的無線多通道使用與波導管方法進行設計，以期望能提升晶片整體效能。而在最後，我們使用 OMNeT++為基礎的網路晶片模擬器來模擬我們的方法。 | zh_TW |
| dc.description.abstract | As the technology progress, CPU performance doubles every 18 months. Now the development of VLSI is faced with physics limitation. For past years, CPU performance is increased by raising up CPU frequency. But by the effect of physics limitation, CPU frequency cannot be grown unlimitedly. Therefore, the number of cores on a chip is increased from single core to multi-core. As the number of cores on a chip grows, the computer network concept has been introduced for the architecture of network-on-chip (NoC). More specifically, NoC provides packet switching based communications in order to connect components on the chip. In the meantime, the introduction of RF (radio frequency) interconnect brings the new opportunity for high data rate, low latency and low power consumption for millimeter range on-chip communications for next chip generation.
In this thesis, we focus on the multi-channel wireless routing algorithm for the two-dimensional and/or three-dimensional NoC structures among processor cores and inducts the new waveguide design to enhance performance. We also study the design of internal routing protocols with channel allocation using Group-Based strategy. This design allows the chip cores to accomplish packet switching by means of using wireless multi-channel grouping and waveguide solution. Finally, we use OMNeT++ based NoC simulator to simulate the system performance. | en_US |
| DC.subject | 無線路由演算法 | zh_TW |
| DC.subject | 多通道 | zh_TW |
| DC.subject | 分群方式 | zh_TW |
| DC.subject | 三維晶片網路 | zh_TW |
| DC.subject | 無線晶片網路 | zh_TW |
| DC.subject | 光學晶片網路 | zh_TW |
| DC.subject | Wireless routing algorithm | en_US |
| DC.subject | Multichannel | en_US |
| DC.subject | Group-Based method | en_US |
| DC.subject | 3D Network-on-chip (NoC) | en_US |
| DC.subject | Wireless Network-on-chip (WNoC) | en_US |
| DC.subject | Optical Network-on-chip (ONoC) | en_US |
| DC.title | 以分群為基礎之3D無線與光學網路晶片頻道存取方法 | zh_TW |
| dc.language.iso | zh-TW | zh-TW |
| DC.title | Group-based Channel Access Scheme for 3D Wireless and Optical Network-on-chip | en_US |
| DC.type | 博碩士論文 | zh_TW |
| DC.type | thesis | en_US |
| DC.publisher | National Central University | en_US |