Teaching Awards and Experience 

When I joined McGill, there were no courses in 
high-performance computer architectures and parallel 
processing systems. I introduced and developed a set of 
new courses (308-505,308-605,308-622) on computer
architectures and parallelizing compilers. These 
courses have been consolidated and improved over
the period of time, forming a core for students who are 
interested in the related subject areas. For example, 
the enrollment of 308-505 has trippled over the years. 
I have also taught a variety of new graduate
seminar courses as listed below. 

    

    

  1. outstanding teaching award nominations:

      

    1. nomination for the Engineering Class of 51' 
Award for Outstanding Teaching (1988);

      

    2. nomination reconsideration for the Engineering 
Class of 51' Award for Outstanding Teaching 
(1989);

      

    3. nomination for the Engineering Class of 51' 
Award for Outstanding Teaching (1990);

      

    4. nomination reconsideration for the Engineering 
Class of 51' Award for Outstanding Teaching 
(1991);

      

    

    

  2.  New courses introduced:

      

    1. 308-505:  Topics in High-Performance Computer 
          Architecture 
This course examines the basic principles and
modern techniques used in the design of
high-performance computer architectures. The
course stresses the concepts which are embodied
in many high-performance computer architectures,
ideas which will continue to apply in the
future in spite of a rapidly changing
technological environment.  The primary
objective of the course is to show how
effective designs use technology cleverly
to achieve balanced, efficient structures, well
matched to the class of problems they overcome
Both qualitative and quantitative approach
has been covered in the course.

      

    2. 308-605:  Topics in Parallel Computer 
          Architecture 
This course studies the basic principles in
parallel computer architecture and programming 
issues. The course stresses the concepts and
principles of parallel processing, and its rela- 
tion with the underlying computer architecture. 
The fundamental issues of von Neumann style -
multiprocessing are addressed. Novel architec- 
ture models such as dataflow as well as multi-
threaded architecture models are discussed.Both 
theoretical and pragmatic issues in designing 
parallel computer architectures are covered.

      

    3. 308-622:  Advanced Topics in Compilers 
One of the main challenges to future parallel
computer systems is their programmability.Thus, 
we need to study and develop compiling 
methodologies which can effectively perform 
(1) program analysis, and (2) program mapping
(decomposition, scheduling,and data structure 
mapping) to maximally exploit parallelism
with efficiency. This course studies the
contemporary work in compiler design and
implementation for parallel computer systems
such as vector/pipelined machines, superscalar/
superpipelined machines, and SIMD/MIMD
multiprocessor systems. In addition, compiling 
issues for novel architectures with fine-
grain parallelism are discussed.

    

    

  3. New software tools introduced or developed
for teaching:

    

      

    1. the SISAL language, compiler and program 
profiling tools from LLNL;

      

    2. the Stanford AWB Architecture Workbench;

      

    3. the Berkeley/Stanford DLX Architecture
simulation tools;

      

    4. the McGill Compiler and Architecture Testbed;

      

    5. the McGill Dataflow Architecture Testbed;

      

    6. an instrumented cache trace simulator;

      

    7. various benchmark suits for architecture/
compiler studies: SPEC, LINPACK, Whetstone, 
Drystone,Livermore Loops, NAS, Spice, GCC etc.;

      

    8. The MTA testbed

      

    9. The SITA tool set

      

    10. The EPPP tool set

    

    

  4. Graduate seminar courses taught:

      

    1. 308-762A:  Topics in Languages and Compiler 
           Design for High-Performance 
           Computing (Fall 1989,1990)  
This course was introduced and developed in the
1989 and 1990, and is now called 308-622.

      

    2. 308-764A:  Introduction to Supercomputer: 
           Architectures, Programming and
           Applications (Fall 1988) 
A substantial contents of this seminar course
has later been subsumed in 308-605 which began 
in 1989.

      

    3. 308-766B:  Topics in Functional Programming 
           and Architectures (Spring, 1988) 
In this course, the students are guided 
in intensive reading of a well selected set of
contemporary literatures in the field of
functional programming, dataflow languages,
and architectures.  Class debates are organized
around the two fundamental issues in multi-
processor computers: the memory latency and 
synchronization overhead.  Novel architecture 
models are discussed which address the two 
fundamental issues. The programmability of 
these architectures are also discussed.

    

    

  5. The ACAPS laboratory

I have established the Advanced Computer 
Architecture and Program Structures Laboratory
since 1988. In addition to performing first class 
research, one important objective of this 
laboratory has been to facilitate the teaching of
the computer architecture and compiler courses,
and training of graduate research and teaching 
assistants. The new courses and software
tools described above depend directly on 
this laboratory. The laboratory is now equipped 
with various workstation. We have a wide variety
of research and teaching software installed,
and a number of my best graduate students have 
actively participated and contributed to teaching.