透過您的圖書館登入
IP:3.21.76.0
  • 學位論文

增進積體電路設計可靠度之研究方法

Study on Enhancing Reliability of Integrated Circuit Designs

指導教授 : 黃婷婷
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


隨著超大型積體電路的發展,現今的單一晶片可以容納數百萬個電晶體並且保持著相當小的體積,這種技術可以幫助更多更複雜的設計得已實體化。然而,由於使用者在電子產品的效能及體積上的要求漸趨嚴苛,系統的可靠性已經變成重要的課題,因此,我們必需要針對高效能的可靠性設計提出有效的解決方法。 當製成的演進進入了深次微米的階段,有一些特殊的現象使得可靠性的議題變得更具有挑戰性,這些現象如串音的效應,溫度過高的問題,以及電流電阻壓降的效應,都在深次微米的階段更具影響力,因此,在本篇論文中,為了維持系統的可靠性,我們將針對這三種現象分別提出個別的解決方法。 為了處理串音的問題,我們提出了一個匯流排架構來移除最嚴重的串音現象,進而改進匯流排的效能。主要的想法來自於計算機結構中預取機制的特性,在預取機制下,指令或資料的提取速度永遠高於指令或資料的完成速度,因此,利用預取的這個特性,我們提出的方法幾乎不會有效能上的損失。 為了處理系統溫度過高的問題,我們提出了利用編譯器來降低極高溫度點的方法。在超長指令集計算機架構上,需要編譯器的幫助指令排程,來達成指令之間的平行處理,因此,這個研究主要是建構在超長指令集計算機架構上。在這?我們提出了兩種方法,第一種方法是暫存器連結,同時考慮空間及時間上的溫度資訊來達到平衡暫存器之間的溫度。而第二種方法則是指令排程,利用一個指令排程的演算法,使可以進行前饋機制的指令個數增加以減少存取暫存器的次數。 為了處理電流電阻壓降的問題,我們提出了一個X填值的方法來減輕在即時性測試的電流電阻壓降效應。在進行即時性測試時,電路會有極為可觀數量的元件同時發生轉換,使得電路上的供電開始不穩,產生所謂的電流電阻壓降效應。這會使電路的效能降低,甚至功能發生錯誤,使電路本身被誤判為錯誤。故我們需要設法降低在即時性測試中的元件轉換率。在這裡我們提出了兩個方法來達到這個目的,分別是考慮電路裡各元件的實際位置以及向後傳遞X值來決定填值。

並列摘要


As feature sizes scale down to deep sub-micron, crosstalk, heat-induced hotspots and dynamic voltage drop (IR-drop) effects have become more and more important. In this thesis, we will study how to enhance the reliability of integrated circuit design by solving the above-mentioned problems. First, to eliminate the crosstalk effect, we propose a de-assembler/assembler technique to eliminate undesirable crosstalk effects on bus transmission. By taking advantage of the prefetch process, where the instruction/data fetch rate is always higher than the instruction/data commit rate, the proposed method incurs almost no penalty in terms of dynamic instruction count. Second, to remove the heat-induced hotspot, we propose a static thermal management technique at compiler level. Two techniques are proposed. The first one is to balance the temperature of the register file by taking both spatial and temporal thermal information into consideration during register binding. The second one is to improve forwarding methods including forwarding-aware architecture and instruction scheduling to reduce the access count of register file. Finally, to reduce the dynamic voltage drop (IR-drop) effect, we propose an X-filling approach during at-speed test. In this work, we take the spatial information into consideration in our approach and we perform a more efficient backward-propagation X-filling method as compared to previous work using forward-propagation.

參考文獻


[2] R. Arunachalam, E. Acar and S. R. Nassif, "Optimal Shielding/Spacing Metrics for Low Power Design," in Proceedings of the IEEE Computer Society Annual Symposium on VLSI, pp. 167-172, February 2003.
[3] S. K. Wong and C. Y. Tsui, "Re-con?gurable Bus Encoding Scheme for Reducing Power Consumption of the Cross Coupling Capacitance for Deep Sub-micron Instruction Bus," in Proceedings of the Design, Automation, and Test in Europe Conference and Exhibition, pp. 130-135, November 2004.
[5] C. Duan, A. Tirumala, and S. P. Khatri, "Analysis and Avoidance of Cross-talk in On-Chip Buses," Hot Interconnects 9, pp. 133-138, August 2001.
[8] C. Duan and S. P. Khatri, "Exploiting Crosstalk to Speed up On-Chip Buses," in Proceedings of the Design, Automation, and Test in Europe Conference and Exhibition, pp. 778-783, February 2004.
[9] W. A. Kuo, Y. L. Chiang, T. Hwang, and Allen C. H. Wu, "Performance-Driven Crosstalk Elimination at Post-Compiler Level," in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Vol. 26, pp. 564-573, March 2007.

延伸閱讀