Moving AI-driven data between tools and fabs

By Junko Yoshida for OJO & YOSHIDA REPORT, in collaboration with John West, Director of Semiconductor Manufacturing Test & Equipment at Yole Intelligence, part of Yole Group – As SEMI works on the Smart Data-AI initiative, Teradyne embraces “Open Architecture,” rolling out a real-time analytics product for testing that can cleanly interface with the rest of the design and manufacturing flow.

Using AI as a tool is an idea broadly embraced in the semiconductor industry. But putting AI into practice needs validated data easily shared within the flow of design, testing and manufacturing. Stakes are high for an industry that must now forced to handle data coming from many sources in different forms.

The semiconductor industry’s quest for consistent data is foremost in discussions at the Semicon West and Design Automation Conference this week in San Francisco. (…)

In today’s semiconductor market, “test times for some devices are increasing while the costs of test are trending upwards,” observed John West, director of semiconductor manufacturing test & equipment at Yole Intelligence. “The additional test time, particularly for wafer level burn-in and system level test, allows for significant amounts of data to be collected and analyzed.”

Chipmakers like Intel, Apple and Qualcomm are using their own AI programs to work through the data to help better identify chips that pass standard tests but may still fail in the field.

West believes that “chipmakers like Intel, Apple and Qualcomm are using their own AI programs to work through the data to help better identify chips that pass standard tests but may still fail in the field.” Assuming that they are using the data structures/strategies that are customized for their own usage, West noted that there is an opportunity now to create standards that will enable the whole industry to move forward more efficiently.

… Read the full article here.

Source: https://ojoyoshidareport.com/