Your Research Proposal
Your research proposal is a great opportunity for you to demonstrate your knowledge and understanding of computer architecture. It is also a test of your ability to identify, motivate and describe an interesting and novel research direction. For this reason it is essential that your proposal be an original piece of work. I am very happy to comment on a draft, but I won't tell you what to write!
A good research proposal will take time to prepare and is certainly something you should put a lot of effort into. Be careful to ensure that you make a detailed argument for why your proposed research direction is interesting and timely. Try to be as precise as possible when you write, especially when motivating your proposed research and outlining your idea. Be careful not to assume that the reader knows what you are thinking. Try to identify the important research questions, rather than describing a large number of well known solutions.
A good research proposal will look into the future and consider tomorrow's problems, not today's. If you have indentified an active research area, try to think what the next important challenge in that area might be. Focus on developing and exploring your ideas and proposed direction rather than discussing current trends or related work. One last pitfall to avoid is to simply paraphrase material collected from these web pages, this tells me very little about how imaginative you are.
Useful Links
More information regarding your research proposal can be found here.
Guidance for graduate students on writing a research proposal
Keynote talks from the top computer architecture conferences (ISCA, MICRO, HPCA, ASPLOS, etc..) may also be a good source of inspiration.