TITLE:

Towards More Robust and Reliable VLSI Circuits: A Data Mining Perspective

DATE:

Friday, November 6th, 2015

TIME:

3:30 PM

LOCATION:

GMCS 214

SPEAKER:

Dr. Ke Huang. Department of Electrical and Computer Engineering, SDSU

ABSTRACT:

Millions of Very-Large-Scale Integration (VLSI) devices are fabricated, tested and delivered to the open market every day. As a result, a huge amount of data is being collected from the manufacturing process. For example, high-volume characterization data, yield data, performance test data, etc. are permanently produced and collected. Understanding these data for a given design under a particular technology is a recurring problem in many applications including pre-silicon design validation, post-silicon validation, process monitoring, yield enhancement, etc. However, these data are not being totally explored and understood by process and test engineers nowadays. On the other hand, fabricated devices are not as reliable as we may expect. For example, a measurement value across different dies on the same wafer may vary due to process variations, which would lead to device failure. Random spot defects would also result in catastrophic behavior of a device. Other reliability and security issues include counterfeit activities and malicious hardware modification of the device due to design outsourcing, migration of fabrication foundries to low-cost areas across the globe. In order to verify the functionality of integrated circuits under these reliability risks, we need to test them, which is not an easy task either, because 1) the testing time is long and 2) the automatic test equipment used to perform these tests is very expensive. As a result, we need to verify the functionality and robustness of devices in a time efficient and cost effective manner. In this talk, we will show the use of machine learning and data mining techniques to explore these IC manufacturing data in an effort to solve the aforementioned reliability and robustness issues that we are facing today.

HOST:

Dr. Jose Castillo

DOWNLOAD: