Scribd
Upload
40 views2 pages

Sylabicurrenttrends

This course explores recent innovations in computer architecture and how the compiler and hardware can work together to optimize performance. Topics covered include multithreaded architectures, speculative execution, power-efficient design, value prediction, feedback-directed optimization, and runtime optimization. Students will gain an understanding of design trade-offs through comparing hardware and compiler techniques. Performance is important but not the only focus; improving other metrics like reliability will also be discussed. Evaluation is based on paper summaries, presentations, and a class project.

Uploaded by

ric_4_gen
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
40 views2 pages

Sylabicurrenttrends

This course explores recent innovations in computer architecture and how the compiler and hardware can work together to optimize performance. Topics covered include multithreaded architectures, speculative execution, power-efficient design, value prediction, feedback-directed optimization, and runtime optimization. Students will gain an understanding of design trade-offs through comparing hardware and compiler techniques. Performance is important but not the only focus; improving other metrics like reliability will also be discussed. Evaluation is based on paper summaries, presentations, and a class project.

Uploaded by

ric_4_gen
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

c cc


 c
 
 
 
c
   
News | Syllabus | Lectures | Schedule | Summaries | Projects | Useful






 !


"#$%$ 

c 
&
"" 

This course explores the motivation and design issues behind recently emerged innovations in
computer architecture and investigates how the compiler and the hardware can work in tandem to
fully exploit the potential of each innovation. By comparing and contrasting hardware-base and
compiler-based techniques in architecture design, the students will gain an in-depth
understanding of the design trade-offs in modern architecture. Performance, although very
important, is not the only criteria in optimizing computer systems, this course will also discuss
recent proposals on improving the non-performance metrics in computer systems.

This course will cover but is not limited to the following topics:

˜Ê ultithreaded Architectures
˜Ê Speculative execution
˜Ê Power-efficient design
˜Ê alue prediction
˜Ê -eedback-directed optimization
˜Ê Duntime optimization
˜Ê Alias analysis
˜Ê Architectural support for reliable system

$

'


[raduate standing. Students should be familiar with basic computer architecture concepts such
as pipelining, caches, cache coherence protocol, and branch prediction. Exposure to classic
compiler optimization techniques such as dataflow analysis, register allocation and alias analysis
is a plus.

c 



There is no textbook. The course would be based on research literature and recent publications in
architecture and compiler conferences such as ASPLOS, PLDI, ISCA, icro and C[O.

c 
( )
The course is organized as a seminar where we discuss a different topic every week. Two to
three papers will be assigned to each topic. All students are expected to read the papers and
submit part 1 of the summary before class. One student is responsible for presenting the paper
and leading the discussion during class. Part 2 of the summary is due 2 days after class. All
summaries will be available online. There will also be a written assignments and a class project.

[  *

Paper
30%
summaries
Presentations
30%
and discussion
Project 40%

    

Antonia Zhai (zhai at cs dot umn dot edu), 6-189 EE/CSci Building, 626-1285

÷  

      

  

÷ 

   
  

   



You might also like