以模式配對及二元決策圖從序向真值表重建單元邏輯電路之功能

Abstract

在數位設計中，未知值（X）的存在可能會導致誤解，導致X悲觀主義和X樂觀主義，這可能會誤導設計人員並導致不必要的調試工作。為了緩解這些問題，對使用者定義基元進行全面的引腳識別的需求日益明顯。我們提出了一種結合啟發式方法、結構感知技術和二元決策圖 (BDD) 的方法。

我們的方法旨在識別和理解設計中引腳/輸入的行為，從而實現更準確的分析和最佳化。具體來說，我們採用啟發式模式匹配方法來區分閂鎖、正反器和具有多個選擇/啟用功能的單元。

此外，我們探索了 二元決策圖 的應用來有效地表示和操縱設計的邏輯，促進徹底的分析和最佳化，從而提高工業單元庫的辨識度。大規模商業設計的實驗結果證明了我們提出的方法在解決引腳功能識別問題方面的有效性和效率，從而提高了數位設計的可靠性和正確性。

In digital design, the presence of unknown values (Xs) can lead to misinterpretations, resulting in X-pessimism and X-optimism, which may misguide designers and lead to unnecessary debugging efforts. To mitigate these issues, the need for a comprehensive pin recognition for user-defined primitives is increasingly apparent. We proposed an approach that combines heuristic methods, structure-aware techniques, and the utilization of Binary Decision Diagrams (BDDs).

Our approach aims to identify and understand the behavior of pins/inputs within the design, enabling more accurate analysis and optimization. Specifically, we employ heuristic pattern matching methods to distinguish between normal latches, flip-flops, and cells with multiple select/enable functions.

Furthermore, we explore the application of BDDs to efficiently represent and manipulate the logic of the design, facilitating thorough analysis and optimization, which improves the recognition of industrial cell libraries. Experimental results on commercial designs demonstrate the effectiveness and efficiency of our proposed methods in addressing pin function recognition problem, thereby improving the reliability and correctness of digital designs.