Special Seminar for Computer and Information Sciences - Yamin LI
- 2012 年度版 (2013年度版準備中)
Instructor
Goal and Theme
High-performance CPUs and supercomputers
Abstract
All design details, including high-performance processor, interconnection networks, collective communication algorithm and Fault tolerant routing algorithms, will be taught in the course. Students will search and present published papers and develop their own projects related their research theme.
Schedule
通年
| 回 | テーマ | 内容 |
|---|---|---|
| 1 | Verilog HDL | Verilog HDL |
| 2 | Verilog HDL | Verilog HDL |
| 3 | Multithreading | Multithreading and thread selection algorithms |
| 4 | Multithreading | Multithreading and thread selection algorithms |
| 5 | Multicore | Multicore and cache organizations |
| 6 | Multicore | Multicore and cache organizations |
| 7 | Trace-driven simulation | Benchmark, Trace-driven and execution-driven simulations |
| 8 | Trace-driven simulation | Benchmark, Trace-driven and execution-driven simulations |
| 9 | Interconnection networks | Interconnection networks the top_10 supercomputers used |
| 10 | Interconnection networks | Interconnection networks the top_10 supercomputers used |
| 11 | Topological properties of IN | Degree, diameter, average distance, bisection bandwidth |
| 12 | Topological properties of IN | Degree, diameter, average distance, bisection bandwidth |
| 13 | One-to-all broadcasting | One-to-all broadcasting on hypercube |
| 14 | One-to-all broadcasting | One-to-all broadcasting on hypercube |
| 15 | All-to-all broadcasting | All-to-all broadcasting on hypercube |
授業外に行うべき学習活動
Read papers and develop HDL codes or develop routing algorithms
Materials
None
References
Online papers
Evaluation Method
Developed project and presentations
情報機器使用
Bring note-PC to the lecture
前年度の授業改善アンケートからの気づき
None