PS求打磨!!!

avatar 70918
vito
1986
12
[align="left"]I am fond of being aprofessional practitioner in electrical engineering, especially in field of digital design. I have been concerning on this area since I enrolled inmy university and I was greatly impressed by it. The first glimpse of the fieldof computer engineering is through an open course offered by MIT from iTunesUon electronics. Along with the accumulation of knowledge in this field, I foundmyself particularly engaged in the study of how to “tape out” a delicate designof hardware system, from the bottom layer of abstractions of transistors to thelogical design of structures of Soc and synthesis techniques and methodology. Digitaldesign is exactly what I am looking for.[/align][align="left"]
[/align][align="left"]With the determination to propel me, I worked hard to take extracourses besides mechanics such as Digital Circuit, Synthesis Design ofIntegrated Circuit and Signals and Systems etc. While taking the courseSynthesis Design of Integrated Circuit, I had difficulty initially inunderstanding the idea of using more buffers to disintegrate logic in order tosolve the problem of clock skew and their application in clock tree. I wroteseveral programs in Verilog, including an asynchronous FIFO memory which led toa perspective in comparison with mechanical systems: a signal is like a tinyball tossed out. If it were to pass an obstacle, it has to reach a certainheight, which corresponds to a signal with certain strength and sharpness,before reaching the obstacle. In the process, using the buffer is like reducingthe sickness of the obstacle (disintegrating logic) and it will restore thesignal behavior. What we need to do is to find the balance between attenuationinfluences and cell placements. This endeavor has provided me with a moredirect understanding of how circuits work and a deeper grasp of the concepts inthe electrical engineering.[/align][align="left"]
[/align][align="left"]I also took an internship inthe Chinese Academy of Railway Sciences and during which I gained a lot ofpractical experience. I was mainly in charge of algorithms implementation usingC++ in Linux. Earlier I had experience of programming in Visual C++, but when Ihad a taste of the Linux platform I immediately transferred myself. From a Linux perspective the computer is morelike a huge file tree so you can check nearly every corner of it as you wish. Ialso accumulated a set of developing environments during that process. The algorithmI focused on was the Marching Cubes algorithm for visualizing scalar field. Ittook me less than one month to get myself familiar with the basic graphics pipelinein detail. I started my prototype rather early to create a navigating scene,after that I made the core algorithms working after debugged a hidden floatingpoint exception using gdb. That whole experiences enriched my practical abilityand get me to a deeper understanding of how a command is compiled and executed andits implicit relationship to hardware settings.[/align][align="left"]
[/align][align="left"]After attending the class inIC synthesis I come to know the world of VLSI and digital design. It is a morechallenging world than I expected yet more amazing and fascinating. So after Ifinished the course I took an internship with the professor in a project abouthigh performance GPU with limited energy and area consumption on mobiledevices. I spent a couple of days trying to understand the computer pipeline notas a user but as a server. During the process of understanding the basicknowledge of pipeline division and graphic vertex and pixel pipelineoperations, an idea came to my mind: to achieve a better performance, if theclipping process is placed before perspective transformations, those dummy datawould not require hardware resources. Bearing this in mind I searched forpapers in architecture pipeline and later talked to my advisor. I tried a smallmodel on an FPGA board but proved to be not that successful. The result and thefeedback from other resources suggest that though a double depth-test may seemto be removing data, that process itself has extra area and elements cost. Whatis worse, the first depth-test is based upon projection relationship whichmeans it could be unsettled. From that experience I learned the balancingbetween different needs and how a design idea grows into a product.[/align][align="left"] [/align][align="left"]I am highly aware of yourschool XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX[/align][align="left"]
[/align][align="left"]还恳请各位批评指正了!!![/align]
  • 3
12条回复