Knowing where one is heading during navigation brings assured happiness. As a student majoring in Computer Software, I began my odyssey four years ago. Now, after the initial mysticism was gradually unveiled, my curiosity remains the same. Indeed, having entered this splendid computer world, I am more than greedy for something new.
From the beginning of my study, my endeavor was fixed on the underlying branches of Computer Science, particularly System Software development. Novel applications on other’s platform may be fruitful, but I think it’s more appealing to act as an independent “manager”. In fact, mathematics, OS, DBMS and modern compiler are all the examples, any breakthrough of which would push forward the whole industry. Individuality is achieved in this unique position.
My paces toward this goal are always steady. As mathematics permeates to the every corner of Computer Science, I am eager to see how it functions. I took courses offered by the Mathematics Department including Mathematical Analysis and Advanced Algebra. The curriculum also covered Discrete Mathematics, Probability & Statistics and Theoretical Computer Science. As supplement to my scope of knowledge, I learn by myself Combination Mathematics and the Science of Programming. This really made a hard period of time, but the harvest was rewarding. I come to understand that even the most irrelevant software disciplines have the origins in common. The importance of Fractured Geometry in Computer Graphics is already obvious. What if a step furthers toward TSP or Bin Packing? Immeasurable. Then came my favorite topics: Operating System, Compiler and Database. I worked hard and derived bits of my own insight. In fact, I was greatly encouraged to find some of my ideas successfully implemented in the corresponding course projects. My final grade is straight “As” in these coursed. In short, although my experience in Computer Science is still limited, I believe its depth is well accessible. As my advisor, Prof. Fang Yu, put it figuratively in one of his lectures: “ It makes no difference whether a hunter captures 5 or 7 rabbits. What counts is he knows how to use his gun.”
I think I can be the qualified shooter now. in my undergraduate years, I have earned various kinds of scholarships, among which were “Peking University Fellowship” and “Excellent Academic Scholarship”. My overall GPA ranks upper 10% among 48 students of the same grade. Because of my satisfactory performance, I was granted the honor of entering the graduate program at Peking University directly, waived of the admission test. In retrospect, my workload is always heavy but it is worth my time of effort. Presently, I have both adequate theoretical understanding and rich programming experience. READY I AM.
Of all the sub-areas of Computer Science, my major interest is parallel processing and the related compiler construction. The terminology of parallel processing came to me when I read an article about digital circuit testing. According to this article, there exist many serial approaches of solving the automatic test pattern generation (ATPG) problem. Though certain method of solving ATPG sometimes proved efficient, the computation time for a method of solving ATPG sometimes proved efficient, the computation time for a complete result is intolerable, the a parallel processing machine is proposed. Using paralleled ATPG method, this NP complete dilemma is ingeniously handled. After reading, I surprisingly realized that parallel processing opened a new window for the computationally intensive problems, which prove intractable for the VON NEUMANN Architecture. You can always first split a specific computational task into several independent divisions, then assign each part to a certain processor. With reasonable partition and proper synchronization, most grueling scientific applications can be reduced to normal scale problems.
With the advent of the multiprocessor architecture, compiler optimization faced a new challenge. A compiler not only has to translate, but it is also responsible for generating parallel-processing loop provides a rich opportunity of exploiting parallelism. Therefore, finding an optimum strategy for loops because a hot research topic. On the whole, parallelism changes our view toward the conventional computer architecture so that the compiler’s construction must make the corresponding modification. Because of this challenging characteristic, I choose them to be my interest of future study.
Yet parallel processing is far from maturity. It has been observed that some sequential algorithms can be easily translated into fast parallel algorithms while some others may not. Is every sequential computation can be done efficiently in parallel? If not, what properties result in such distinction? Another unsettled problem is programming parallel computers now is like programming serial computers in the 1960s. Programmers get locked into a particular manufacturer’s computers. Is it possible for us to design a higher level language for parallel computers that hides the underlying machine architecture from users? Both are crying out for solution.
Without advanced knowledge, I can do nothing but wait. In order to convert my undergraduate study into full play, graduate education is a necessity. From the college guide and my advisor’s recommendation, I am pleased to learn that your department has an academic atmosphere where originality is appreciated and individual potential is explored. It will be the best place for my specialized study. If possible, my main interest in the future will be in the following areas:
Parallel processing, Compiler construction.
Distributed/concurrent systems, Operating system.
Machine learning, Automated fault diagnosis.
As my undergraduate education is very basic, if you feel that I am more suited for some other areas, any of your suggestion would be highly appreciated.
Upon the completion of my study, I would return to Peking University and teach what I’ve learned to the students. Then I shall establish a research center to experiment new ideas. It is my ultimate goal to see the Chinese computer products leading the global market in the next century. After all, the navigation is not over yet!