One billion transistors, One Uniprocessor, One Chip
Alma C. Ryan
GS1145
ITT Technical Institute

One billion transistors, One Uniprocessor, One Chip I read an article from University of Michigan 1997 about one billion transistors, one processor, one chip. It has 5 authors- Patt Y., Patel S., Evers M., Friendly D., and Stark J.. At beginning the authors, lead me to believe they were hesitant of the existent of the one billion transistors, one uniprocessor in one chip. They have comments such as, "A billion transistors on single chips presents opportunities for new levels of computing capability". The other comment was, "The basis design problem is partitioning". However, later on, after giving their suggestions, the authors seemed to agree on the released of the said chip. This article was written in 1997. Few years before the said chip was put in the market. This article discussed both pros and cons of the upcoming one chip. One of the pros was , "the uniprocessor will be high performance". Another pros was, "that this chip will have large trace cache, and large number of reservation stations". Additionally" this chip will also have a sufficient resolution and forwarding logic" another pros that the authors mentioned. But the authors mentioned more cons than pros. Some of the cons discussed were: "one billion transistors falls far short of an infinite number, the highest performance computing system will not fit onto a single chip", according to the authors. Additionally, the authors stated that, "a chip processor does not correctly address the partitioning problem". They further elaborated by saying, " The multiprocessor divides the available transistors among processors on the same chip". Another cons mentioned by the authors that I found interesting was that, "such engine could never run