大量診斷與偵錯在辨別由製程、設計以及測試問題造成系統性測試故障的系統性缺陷時扮演非常關鍵的步驟。 然而診斷工具時常因測試通過模擬錯誤、多重缺陷診斷以及壓縮失真的問題導致診斷解析度不良。 在這篇論文中,我們提出利用測試故障群聚與重組的技術以提升診斷解析度。我們用兩個先進製程的業界設計晶片測試故障來展示技術的成效。我們的技術在兩個設計上的平均診斷解析度分別提升了9.42倍與575.25倍。我們的技術同時減緩壓縮失真並成功辨別出所有錯誤的掃描單元。靜態時序分析與診斷性測試驗證了我們的技術辨別出的系統性缺陷。我們的技術可以基於現有的診斷工具實現並僅有低於1\%的額外執行時間。實驗數據顯示我們的技術提升良率學習品質並且發現了一個無法被傳統大量診斷發現的設計相關缺陷。
Volume diagnosis and debug play a key role in identifying defects related to process, design, or test issues and thus causing systematic test failures. However, diagnosis tools often suffer from poor diagnosis resolution due to issues of Test Pass Simulation Fail (TPSF), multiple defect diagnosis, and compression aliasing. In this thesis, we propose techniques to improve diagnosis resolution by test failure clustering and reorganization. The effectiveness of our techniques is demonstrated on test failures of two industrial designs in cutting-edge process nodes. Our technique improves the average diagnosis resolution on the two designs by 9.42x and 575.25x, respectively. Our techniques also mitigate compression aliasing and successfully identify all failing scan cells. Static timing analysis and diagnostic testing have verified the systematic defects identified with our techniques. Our techniques can be implemented using existing commercial diagnosis tools with runtime overheads below 1%. Our techniques are shown to enhance yield learning quality and uncover a design-related defect that cannot be found with traditional volume diagnosis.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。