Lectures
You can download the lectures here. We will try to upload lectures prior to their corresponding classes.
-
芯片发展史与AI芯片体系结构简介
Summary: 课程情况介绍，简要介绍芯片发展历程，以及简要介绍前沿AI芯片体系结构
[slides] [supplementary slides]
Suggested Readings:
- Online Verilog Exercise
- MIPS Website: The official website of MIPS
- MIT Eyriss: The official website of Eyriss, a hardware CNN accelerator developed by MIT
- RISCV: The official website of RISCV, a simplified instruction set developed by UC Berkeley
- Software Optimization: Introduce common methods for software optimization to accomondate hardware
- Compiler Optimization: 编译器优化相关介绍
-
电路基础-1 晶体管与数字电路设计
Summary: 晶体管原理及数字电路基础，为后续学习更复杂的电路结构做准备
[slides] [supplemantary slides]
Suggested Readings:
- Online Verilog Exercise
- 数字集成电路设计原理: 包含静态逻辑动态逻辑等内容
- CMOS Delay Model: Slides given by UIUC ECE342
- Digital Integrated Circuits - A Design Perspective: A reference book for digital integrated circuits deisgn.
- Verilog HDL高级数字设计: Verilog参考书籍
- Logical Effort - 1: An introdcution to logical effort concept and calculations
- Logical Effort - 2: Slides about logical effort by UT Austin
-
电路基础-2 时序电路、芯片的物理设计与验证
Summary: 介绍时序电路、状态机以及芯片的设计流程，涵盖从前端、后端到最后流片的整体过程
[slides] [supplementary slides]
Suggested Readings:
- Online Verilog Exercise: A website for verilog exercise
- FSM Supplementary: 介绍有限状态机与常见的状态机代码格式
- 乘法器设计: 乘法器设计参考教材
- PGK Adder Tree: Slides for PGK adder tree design
- 时序分析: 时序分析（STA）教材中文翻译版，可以参考第一章第二章中的基本概念
- STA basics and examples: Basic concepts in STA (including setup time, hold time, skew, jitter and etc.)
- SRAM Wiki: Wikipedia page for SRAM
- DRAM Wiki: Wikipedia for DRAM
-
指令集与流水线设计
Summary: 介绍指令集ISA基础与流水线CPU的设计原理
[slides]
Suggested Readings:
- Instruction Set: Appendix A in the textbook
- Pipelining: Appendix C in the textbook
- MIPS ISA: MIPS ISA introduction
-
数据/控制冲突以及处理机制
Summary: 介绍流水线CPU中的数据/控制冲突与处理方法
[slides]
Suggested Readings
- Pipelining: Appendix C in the textbook
- Hazards: P150 - P156 in the text book
-
指令动态发射原理
Summary: 介绍乱序执行指令动态发射原理，以MIPS R10K为例
[slides]
Suggested Readings
-
Instruction Level Parallelism & Branch Prediction: Textbook Chapter 3
-
OoO Arch Slides 1: Slides about OoO Execution
-
OoO Arch Slides 2: Slides about OoO Execution from Stony Brook
-
-
分支预测与超标量设计
Summary: 介绍分支预测技术与超标量技术
[slides]
Suggested Readings
-
Branch Prediction: A Slide about Branch Prediction at CMU
-
Super Scalar: Lecture Recording at ETH
-
Memory Hierarchy: A Slide about Memory Hierarchy.
-
-
多级缓存与缓存一致性
Summary: 介绍多级缓存结构与Cache Coherence相关知识
[slides]
Suggested Readings
-
Memory Hierarchy: Lecture about Memory Hierarchy
-
Cache Coherence: Cache Coherence Lecture
-
-
缓存一致性与预读取
Summary: 介绍缓存一致性、预读取、虚拟缓存与多线程等相关概念
[slides]
Suggested Reading
-
Cache Coherence: Cache Coherence Lecture
-
Pre-fetch: A slide on memory pre-fetch
-
Multi-threading: An introduction to multi-threading with CPU
-
-
-
人工智能加速器架构I
Summary: 介绍人工智能加速器架构
[slides]
Suggested Readings
- AI Accelerator: A lecture about AI acclerator deign methodology
-
软硬件协同设计
Summary: 介绍前沿软硬件协同技术，如量化剪枝蒸馏等
[slides]
Suggested Readings
- Bit Fusion: Bit-Level Dynamically Composable Architecture for Accelerating Deep Neural Network
- ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantization
- OliVe: Accelerating Large Language Models via Hardware-friendly Outlier-Victim Pair Quantization
- EIE: Efficient inference engine on compressed deep neural network
- SNAP: An Efficient Sparse Neural Acceleration Processor for Unstructured Sparse Deep Neural Network Inference
- NV Sparse TensorCore
-
-
未来AI芯片发展趋势&Guest Lecture
Summary: 总结学期课程，介绍未来AI芯片发展趋势，华为朱晓明老师Guest Lecture
感谢朱晓明老师！
