一、課程說明 (Course Description)
We will discuss several advanced concepts of computer architecture for modern machines. The focus will be
on both single-thread and multi-core performance. The parallelism of instruction execution, data/vector, and
threads are examined in details. The concepts will be reinforced through simulation assignments. Students
will perform hands-on simulation and analysis of several architecture techniques. It is recommended that
students have taken undergraduate computer architecture course.
二、教科書 (Textbooks)
“Computer Architecture --- A Quantitative Approach” by John Hennessy and David Patterson, 6th Edition,
2017, Morgan Kaufmann
三、參考書籍 (References)
* Synthesis Lectures on Computer Architecture, Springer
四、教學方式 (Teaching Method)
Lectures.
五、教學進度 (Syllabus)
Weeks Subject
1 Introduction
2 Cache
3 DRAM organization
4 Cache coherence
5 Memory consistency
6 Network on Chip
7 Instruction level parallelism (ILP) and static scheduling
8 Dynamic scheduling/Midterm Exam
9 Hardware speculation
10 Branch prediction
11 Vector machine
12 Vector machine
13 SIMD
14 GPU
15 Domain-specific architectures
16 Final Exam
架構模擬器實驗平台將另提供線上講義和操作文件,以補充17/18週內容和進度。
We will provide lectures and related lab notes of GEM5 simulator for students to complete the 17th and 18th
week workload.
六、成績考核 (Evaluation)
* Homework and simulation assignments 40%
* Midterm Exam 30%
* Final Exam 30%
七、課程先修 (Prerequisite)
計算機結構。
八、使用 AI 的規則 (Rules of AI Tool)
(2)有條件開放,請註明如何使用生成式AI於課程產出
作業(問題回答和程式)不可以抄襲,包括使用AI生成工具,報告中文字可使用AI生成工具修飾,但需註明使用範圍和
方式。
Plagiarism (including AI generation tool) is not allowed for homework assignment (problem solution and
programming). Texts in a report can be polished with AI tool by noting the parts and usage.
九、可連結之網頁位址 (Course material and streaming link)
* eeclass.nthu.edu.tw
Lecture notes, homework assignment, QA.
* Teams: EE6455高等計算機結構 (2023 Fall)
Streaming and recording.
https://teams.microsoft.com/l/team/19%3a9etEdXK0VkSxu6WR2crBQCOnJfdZozRrjXxi53NvV4A1%40thread.t
acv2/conversations?groupId=d53cfd6b-2a5e-4fdb-bf15-f4bdfb8f8174&tenantId=6c3bc511-43c7-4596-
baeb-2335c69c41f1