Semiconductor manufacturers continuously seek to improve yield and reduce test time with adaptive test techniques. Most adaptive test today involves a long feedback loop with test results moved outside of the test cell and actions fed back at a later time, possibly several lots later. These situations will continue to exist in cases where either human supervision is necessary or analysis is done with data from several test cells. However, with advancements in machine learning and analytics, many companies are starting to develop adaptive test algorithms and techniques that could be used with data on a single test cell from within a lot and possibly even during a touchdown. There are legitimate concerns to deploy these on host controllers due to possibility of using compute resources and impacting IG-XL throughput. Hence the need for an edge compute device.

The UltraEDGE was designed to provide a flexible and secure environment to run analytics within a tests cell. This paper will demonstrate how a user can integrate the UltraEDGE for adaptive test within the test flow as well as across the touchdowns.