This presentation will begin with a brief background on GPU (Graphic Process Unit) and its evolution in the past two decades along with the advance of Moore's law. It would then lead to the discussion of Moore's law in present and "More Moore's". As NVIDIA is transforming to a visual computing company participating in the growing market of Deep Learning, Artificial Intelligence, and Autonomous Driving, the presenter would use company's latest flagship made by 16nm FinFET technology to illustrate the significance of the nano-manufacturing technology. He will describe the technological challenges to manufacture such a complicated chip in order to achieve performance, perfection, and precision required for the product. In addition, "More than Moore" will be discussed in the context of heterogeneous integration of GPU and HBM (High Bandwidth Memory) on Si Interposer for high performance computing. At last, some perspectives on semiconductor industry in the foreseeable future will be shared.
John Y. Chen has been in the semiconductor industry for 30 some years ranging from IDM (Integrated Device Manufacturer) to Foundry to Fabless. He currently serves as the vice president of technology and foundry management at NVIDIA Corporation. Prior to that, Dr. Chen held executive positions at FlexICs Inc., TSMC, WaferTech LLC, and Cypress Corporation. Earlier in his carrier at Hughes Research Lab and Xerox Palo Alto Research Center, he has contributed in CMOS technology with more than 100 research papers and a book on "CMOS Devices and Technology for VLSI," published by Prentice Hall. He has taught at Chiao Tung University and Santa Clara University. He was elected to the IEEE Fellow in 1992 for "leadership in and contributions to CMOS device and process technology". Dr. Chen holds a B.S. in E.E. from National Taiwan University, an M.S. in E.E. from University of Maine, a Ph.D. in E.E. from UCLA, and a Master degree from the UCLA Executive Engineering Management Program. .