Learn · NPTEL

Course10610800360 lectures

Computer System Optimizations for ML

Prof. Sumit K. Mandal · IISc Bangalore

Start with lecture 1 At NPTEL

About

Increasing complexity of machine learning (ML) algorithms have necessitated the emergence of specialized computer systems. Some ML algorithms are executed at edge, some are executed on the cloud. In this course, we will delve into different computing kernels of both training and inference of ML algorithms and see how they can be efficiently computed. Specifically, we will cover different system optimization techniques for convolutional neural network (CNN), large language models (LLMs) and Graph Neural Networks (GNNs).

Intended audience: Final year UG, Final year PG, PhD

Prerequisites: Computer Organization

Lectures

Week 1 Introduction

  1. Introduction to the course YouTube
  2. Neural Networks YouTube
  3. Introduction to Convolutional Neural Network - 1 YouTube
  4. Introduction to Convolutional Neural Network - 2 YouTube
  5. Convolutional Neural Network-1 YouTube

Week 2 Convolutional Neural Network

  1. Convolutional Neural Network-2 YouTube
  2. Convolutional Neural Network-3 YouTube
  3. Canonical CNN Architectures - 1 YouTube
  4. Canonical CNN Architectures - 2 YouTube
  5. CNN Computational Optimization Techniques YouTube

Week 3 CNN Accelerators

  1. Dataflow for CNN Execution - 1 YouTube
  2. Dataflow for CNN Execution - 2 YouTube
  3. CNN Accelerators - 1 YouTube
  4. CNN Accelerators 2 YouTube
  5. CNN Accelerators 3 YouTube

Week 4 CNN Optimization

  1. CNN Accelerators 4 YouTube
  2. CNN prunning - 1 YouTube
  3. CNN pruning - 2 YouTube
  4. Parallelism in CNN - 1 YouTube
  5. Parallelism in CNN - 2 YouTube

Week 5 LLMs

  1. Parallelism in CNN - 3 YouTube
  2. Parallelism in CNN - 4 YouTube
  3. Parallelism in CNN - 5 YouTube
  4. LLMs-1 YouTube
  5. LLMs-2 YouTube

Week 6 LLMs

  1. LLMs-3 YouTube
  2. LLMs-4 YouTube
  3. LLMs-5 YouTube
  4. LLMs-6 YouTube
  5. LLMs-7 YouTube

Week 7 LLMs

  1. LLMs-8 YouTube
  2. LLMs-9 YouTube
  3. LLMs-10 YouTube
  4. Distributed Service YouTube
  5. Fault Tolerance YouTube

Week 8 Model Checkpointng

  1. Fault-Tolerant Training YouTube
  2. Fundamentals of Graph Neural Networks-1 YouTube
  3. Fundamentals of Graph Neural Networks-2 YouTube
  4. Hardware Acceleration of Graph Neural Networks-1 YouTube
  5. Hardware Acceleration of Graph Neural Networks -2 YouTube

Week 9 In-Memory Computing(IMC)

  1. ISSAC, Pipelayer YouTube
  2. Pipelayer[contd.], Atomlayer, Simulation platform YouTube
  3. Simulation platform, Area-aware optimization, communication-aware optimization YouTube
  4. Communication-Aware optimization, Achieving minimum communication latency YouTube
  5. Achieving minimum communication latency YouTube

Week 10 GNN Accelerator

  1. Advanced Hardware Architectures for GNN Acceleration YouTube
  2. Communication-Aware Pruning for IMC YouTube
  3. Re-Transformer YouTube
  4. Performance Evaluation of 2.5D IMC-Based AI Accelerators YouTube
  5. Chiplet-Based Heterogeneous Architectures for LLM Acceleration YouTube

Week 11 Performace Characterization of Deep Learning Model

  1. Chiplet-Based Heterogeneous Architectures for LLM Acceleration YouTube
  2. Performance Characterization YouTube
  3. Performance Characterization of Deep Learning Workloads - Part 1 YouTube
  4. Performance Characterization of Deep Learning Workloads - Part 2 YouTube
  5. Performance Characterization of Deep Learning Workloads - Part 3 YouTube

Week 12 NPU and GPU

  1. Introduction to AMD Neural Processing Unit (NPU) YouTube
  2. Introduction to AMD Neural Processing Unit (NPU)[contd.] YouTube
  3. Acceleration of ML operation with GPUs YouTube
  4. Training Large model with limited GPU memory YouTube
  5. Recap of the course YouTube

Syllabus and materials

At NPTEL, as the course publishes them.

Credits and licence

Institution
IISc Bangalore
Department
Computer Science and Engineering
Teacher
Prof. Sumit K. Mandal
Level
Undergraduate
Language
English
Subjects
Computer science, Engineering
Source
NPTEL
Licence
CC BY-NC-SA, as NPTEL states it. Non-commercial use only. NPTEL's terms
Attribution
NPTEL states: "Distributed under Creative Commons Attribution-ShareAlike / CC BY - NC - SA", beside a BY-NC-SA badge.
Who can watch
NPTEL gives no rating. Invisible Network shows it to Tweens profiles and older, judged from what the source says about it. How courses are placed