Young H. Cho (young@myri.com)

Research and Work Experience

Updated on November 1, 1998

UNDERGRADUATE RESEARCH PARTICIPATION

Unlike most students, I began my undergraduate research in my freshman year in UC Berkeley. Professor Ping Ko of electrical engineering department selected me to be his Senior Engineering Aide. During the year I was in the research group, I participated in developing MOSFET modeling simulation software. The research used physical behavioral data from actual MOSFET to compare with sets of data from model equations. These comparisons were used for latest sub-micron MOSFET technology development. (December 1992 to December 1993)

My early research experience and large scale course projects led me to join another computer research group in UC Berkeley, Networks of Workstation (NOW) group. Three reputable Professors in UC Berkeley, David Patterson, David Culler, and Thomas Anderson oversaw the NOW group. Out of which, Professor David Culler became my research advisor. Under his guidence, my research task was to develop a high-speed network (Myrinet) interface board Hewlett Packard GSC bus and its documentation. I eventually learned to use and program with many different PC Board design tools, logic design tools, and languages that were used to design the board (i.e. Powerview, View Logic, Visula, Max Plus 2, VHDL, AHDL, ABEL, and etc.) As a result, I produced a complete network interface board and a low-level test software for the board. Then I compiled GSC+/Myrinet interfacedesign documentation. During this time, I presented my research progress to the NOW group at its weekly meetings. (May 1995 to May 1996)

IMPORTANT TEACHING EXPERIENCE

As early as my second semester in Berkeley, I decided to direct all my efforts to understand computer architecture. At the end of my third semester, I have completed two of the most difficult upper division computer courses, Digital Logic Design and Computer Architecture. My main purpose for taking these classes was to understand the higher concept of computer architecture, which is heart of all computer science techonology. In spite their fast paced instructional schedule, I spent more time to thoroughly understand each concept. Thus, I was allowed to take a graduate-level Computer Architecture course on the following semester. By this time, my advanced class projects reflected my increased level of understanding in computer architecture.

Then Professor Patterson asked me to help him teach his computer architecture course along with other graduate students. So in the Spring semester of 1995, I taught my first college level computer science course. I designed my class sessions to allow the students to understand the concept before explainning the details. As a result of my preparation and effort, my discussion class became very popular, transforming it into the largest section with the highest average score. This result helped me to join the NOW group, soon after. (December 1994 to May 1995)

After six months of research work in NOW group, Professor Culler, asked me to become his Teacher's Aid for his computer architecture course. My increased knowledge in computer architecture improved my teaching techniques. By integrating new ideas that went beyond the scope of the course, I managed all my section groups to design superior projects. Again, my class grew to be the largest class with best projects and scores. (December 1995 to May 1996)

In addition to receiving students' respect, these teaching experiences gave me deeper understanding of the computer architecture by allowing my thoughts to go beyond the class room exercises. The successful results also brought me an honor of receiving the Award for Excellent Student Instruction in Computer Architecture Course (CS152) at my graduation.

SIGNIFICANT CLASS PROJECTS

Other than my research and teaching experiences, I completed four large scale course projects at Berkeley; two courses which strengthened my understanding in computer architecture and two courses that gave me experience in interfacing computer with human.

First set of classes were graduate-level Computer Architecture and VLSI course. In architecture course, I developed a single-chip Massively-Parallel Processor simulator. It collected datas to evaluate the performances of differently configured single-chip multiprocessors. In VLSI course, I designed, simulated, and layed out a Very Long Instruction Word Processor in transistor level.

To build better interfaces to interact with human, I completed two different projects in Mechatronics class. I constructed an intelligent, miniature Kangaroo leg model using microprocessor, touch sensors, artificial muscles, and other pnematic devices to study the superior functionality of the organic muscle. I also constructed a pair of eye glasses that tracked the movement of human eyes, then use its collected information to control external motors. To demonstrate its application, I designed the Accomodation Glasses which automatically focused the lens in the glasses for people who are both far-sighted and near-sighted.

In a graduate-level Computer Vision course, I programed a virtual reality hand that was controlled by the photographic images of different hand-shape. Using various image processing algorithms, the program recognized the hand shapes correctly in real-time.

WORK EXPERIENCE

As soon as I graduated from UC Berkeley, I got to practice my expertise from my research and academic trainning. With Professor David Culler's recommendation, Myricom, Inc. hired me as an engineer and researcher immediately.

Within two weeks of my employment, I began to work on real projects. In first two years, I have completed over twenty internal and commercial projects. The most prominent projects are DARPA contracted projects. My first independent project was to "bring two-level multicomputer to life." For this project, I designed and integrated Automatic Target Recognition system into Myricom's two-level multicomputer prototype. The cost-performance results surpassed other systems, at least, on the order of a magnitude. DARPA inspectors were impressed by the our results generated during the demonstration.

I finished building a more advanced two-level multiprocessor node with two times the memory bandwidth using a new generations of Myricom's proprietary processor by end of February of 1998.

Other than the DARPA project, I completed many board/logic designs; including four commercial network boards, two internal testing boards, and several reconfigurable logic designs. While working on these projects, I wrote my own CAD tools and logic tools to aid my design. I designed the tools in modular blocks, thus making it easy for other engineers to interface with. I have recently finished a 12 layer board design that represents the latest technology in reconfigurable computing system and high-speed network.

My most recent project allowed me to practice all my expertise in the area of producing computer engineering product. I have design a system from hardware all the way up to software. This design allowed, for the first time, for our high speed networks, Myrinet, to travel as far as 2 kilometers. I have designed an hardware and the firmware to convert Myrinet protocol to my serial protocol then placing 1.25 Gigabaud of information on to the fiber optics transceiver. This product is functional, and it will be sold as a commercial product in the last quarter of 1998.