Microprocess Evolution Paper Summary
Requirements, Bottlenecks, and Good Fortune: Agents for Microprocessor Evolution
buzzword
- evolution
- new requirements
- bottlenecks
- good fortune
- a science of tradeoffs
- the tradeoff between high performance and low costs
- the tradeoff between high performance and power awareness
- pipelines
- in-chip cache
- branch prediction
- Specialized Functional Units
- out of order processing
- clusters
- chip multiprocessors(CMP)
- simultaneous multithreading(SMT)
- asynchronous and synchronous units coexisting
summary
This paper introduces the evolution of the microprocessor from 1971 to
2001 and give the opinion about the result of evolution.Author proposes
three things. 1) new requirements; 2) bottlenecks; and 3) good fortune.
All above are collectively called collectively
agents. Whats’ more, the author introduces the past implementations
and its pros and cons, such as pipelining, on-chip caches, out-of-order
processing and so on. Moreover, he discusses the future of
microprocessor and predicts how it will likely be for the future
development, just like the new data path, new materials and so on.
Finally, he concludes with what should be done and what thoughts(the author
mentiones the CAD tools) should be modified at that time.
Strengths
The paper is well structured and proposes three interesting things on the result of microprocessor evolution.
The paper introduces many past implementations and evolutions which widely explain the past implementation and give one who has little preknowledge some important introductions.
The paper predicts the future on what the future would be in the realm of the microprocessor.This gives a brand new advice on how to use the past evolution to predict the future.
Weaknesses:
- The good fortune part is too short that I can\’t figure out what this part means.
Ideas for improvement
- Expand the explanation on how good fortune impresses the evolution and give some more examples.
Lessons learned
This paper gives me a good overview of the timeline of microprocessor evolution. I have a better understanding of why the microprocessor would become like what it likes nowadays. What\’s more, this paper gives me a brand new idea about how to predict the future improvements. The three interesting ideas also impress me how to think about the microprocessor a lot. What\’s more, I think these ideas can be suited to other realms of computer science.