计算机专业留学个人陈述计算机专业留学个人陈述falltheareasofcomputerscience，parallelprocessing fascinates me the most.he term first came tome when read the book omputer rganization & esign：the ardware/oftware nterface as part of myomputerrchitecture course. y teaching me how the performance ofcomputers can be improved through pipelining and parallelprocessing， the book reshaped my thinking in computer science.y lab experience deepened my understanding of parallelprocessing. n my third year， worked on a project with r. omatsu， chief researcher of okyo esearch aboratory and r. oseki， one of r. omatsu&apos&aposscolleagues. heprojectwastointroduce the architecture supports that can allow compilerstoderivemoreparallelismfromprograms. hesearchitecture supports include three designs ， a conditionalexecution mechanism to execute instructions without normalconditionaljumpinstructions ，speculativeexecutionmechanism， and a dynamic memory disambiguation mechanism toexecute memory access instructions simultaneously.sundergraduate students seldom get to be involved in actual1 1 / 3 3research， took full advantage of the opportunity ofworking in the lab. particularly enjoyed the seminarsheld twice a week as part of the lab work.t these seminars ， engaged in in-depth discussions withaster&apos&apossstudents under r.omatsu&apos&aposs seasoned guidance.he lab experience is adding weight and insights to mygraduation paper n mprovement of oftware pipeliningusingoopransformations.nthispaper， willintroduce a translation algorithm using loop interchange andloop skewing. his algorithm can minimize the initiationinterval of loop nest to improve the performance of softwarepipelining.hrough my lab experience and other research， havedevelopedalong-termprofessionalobjective ：theadvancement of the parallel processing technology.hile have strong faith in this technology， also know itslimitations as it is today. ome sequential algorithmscan&apos&apostbeeasilytranslatedintofastparallelalgorithms and most compilers can only distribute works ononeplatform. notherunsettledproblemisthatprogramming parallel computers is now still a manual work.ut even the modern language（like ava， +） can do2 2 / 3 3someautomatictranslationwork.facompilercanbedeveloped to translate and distribute work into differentplatforms， personal computers will be able to work likemainframe computers at a fraction of the mainframes&apos&aposcosts.hiswillmakeitpossibletoconnectallthecomputers in the world and speed up all the processes.o help take the parallel processing technology beyondwhat is capable of today ， have to undertake more advancedstudies. his， believe， can only be accomplished in aquality graduate program like yours. plan to pursue amaster&apos&aposs degree with a concentration on parallelingand distributing compilers， operating systems or computer*works， a combination thereof. ith the trainingthatgoes into such an advanced degree ， think will be ableto take great strides towards fulfilling my professionalobjective. hen do， think will be standing on thecutting-edge of the computer science.3 3 / 3 3