E&CE 720-t4: Project Report
Due: Wed Apr 23 noon. Send PDF by email.
In-depth analysis of a computer architecture theme.
Description: Read three to five papers that all address a
particular theme, then write a paper (maximum 15 pages, single spaced, 11pt
font) analyzing the papers.
- Historical trend: pick a topic from the course (e.g. branch
prediction), find papers that span a range of at least 7 years
and that describe either academic or
industrial perspectives of the topic (e.g. new branch
prediction algorithms or choice of algorithms in industrial
processors), and analyze how the topic has evolved over time.
- Industry vs Academia: pick a topic from the course
(e.g. branch prediction), find several papers that describe
academic improvements to the topic (e.g. new branch prediction
algorithms) and several papers that describe industrial
applications of the topic (e.g. choice of algorithms in
industrial processors). Analyze the similarities and
differences between the academic work and the industrial
- Ideal Theory vs Real Tradeoffs: Pick a topic that matches your area of
research or specialization. Find several academic papers that
propose improvements to this topic. Analyze the real-world
tradeoffs (e.g. area, clock speed, power) that would be needed to
implement the academic algorithms. Analyze how well each paper
does in the analysis of real-world tradeoffs.
Papers should be selected from one or more of the following venues:
IEEE Micro journal, or one of the following conferences: Micro,
ASPLOS, HPCA, ISCA.
Other venues are possible, but be sure to check with me in advance if
you want to use a paper from a different venue.
|Description of papers
|Comprehension of concepts and
relating to course material