| Subject name |
Distributed Computing System |
| Charge teacher |
YAMAGIWA Shinichi |
| Lectures target |
1st,2nd |
Class |
|
| lecture room |
|
Course time |
the former term |
| day・period |
MON２,THU２ |
Unit classification |
select |
| Lecture form |
general lecture |
Unit Count |
2 |
| Matter of prepare |
|
| Notes |
|
| Details1 |
☆Sammery
The main objective of this lecture is to survey the fundamental technologies
for the recent high performance computing, especially the ones of parallel and
distributed computing. Through the class, the audiences will become to explain
the state of the art of the high performance computing world, and also to figure
a blueprint of the future parallel and distributed computing technologies.
This lecture uses English textbooks. This makes the audience to master dedicated
expressions on the parallel and distributed computing field. |
| Details2 |
☆Outline of the lecture
The class treats two themes. The first theme is the "multi-processor system". Another is
the "multi-core system". One of the participants will have responsibility to have a presentation
to explain a part of the theme. Every class will have a different lecturer.
☆Plan and Objectives
First part: "Multiprocessor system and thread-level parallelism"
The following five items will be illustrated.
(1) Perspective of the first part
The participants will become to explain architecture of multiprocessor system and process
thread definitions.
(2) Symmetric shared memory multiprocessor system
The participants will become to explain architecture and performance of symmetric shared memory
multiprocessor system.
(3) Distributed shared memory multiprocessor system
The participants will become to explain architecture and performance of distributed shared
memory multiprocessor system.
(4)Synchronization and memory consistency model
The participants will become to explain mechanism of synchronization among processing threads,
and methods and definitions of memory coherency.
(5) Examples of the systems and Future directions of the field
The participants will become to discuss the future directions using the knowledge learned
in the previous classes.
|
| Details3 |
Second part: "Multicore system and GPGPU"
The following five items will be illustrated.
(1) Introduction to GPGPU and the history
The participants will become to explain advanced architecture of muticore processors and
high performance computing using graphics processing units.
(2)CUDA
Learning introduction to CUDA and the programming methodology, the participants will become
to begin programming application on CUDA.
(3)Applications of GPGPU
Learning experimental application examples on CUDA, the participants will become to explain
which kind of application can fit into the computing mechanism of GPGPU.
(4)CUDA
Learning introduction to OpenCL and the programming methodology, the participants will become
to begin programming application on OpenCL.
(5) Future directions of GPGPU
The participants will become to review the perspective of the GPGPU and also to discuss
the future directions of GPGPU.
|
| Details4 |
☆Textbooks
Several chapters will be picked up from the following three books;
・David A. Patterson and John L. Hennessy,
"Computer Architecture, Fourth Edition: A Quantitative Approach",
Morgan Kaufmann, ISBN:0123704901, September 2006
・David B. Kirk and Wen-mei W. Hwu,
"Programming Massively Parallel Processors: A Hands-on Approach",
Morgan Kaufmann, ISBN: 0123814723, February 2010
・Hubert Nguyen, "GPU Gems 3",
Addison-Wesley Professional,ISBN: 0321515269, August 2007
Note) One week before, the material for the first lecture will be distributed to the participants.
The participants have to review the material in advance.
|
| Details5 |
☆Evaluation
- Using the materials used in presentation, report in the end of
the quarter and mini exams (performed in the beginning of the class),
the grade will be decided.
- The audience who got totally 60% will pass. The materials used in
presentation is 40%. The report in the end of the quarter is 40%.
And the mini exams are totally 20%.
90% or higher: AA
80-89%: A
70-79%: B
60-69%: C
Less than 60%: F
|
| Details6 |
|
| Details7 |
|
| Details8 |
|
| Details9 |
|
| Details10 |
|
|