AI Architecture and Systems Tutorial
This tutorial covers six layers: Microarchitecture → Kernel → Compiler → Architecture → System → Algorithms.
Maintainers: Yufeng Gu and Alireza Khadem
Microarchitecture
How does GEMM execute inside an accelerator?
PE, systolic array, memory hierarchy, dataflow, roofline, and operand delivery.
4 sections · 4 topics
Kernel
Operator Implementation and Performance Optimization
How are operators implemented efficiently?
Layout, tiling, pipelines, matrix instructions, libraries, and device-specific kernel programming.
Planned
Compiler
Compilation and Program Mapping
How does a model become an executable device program?
Graph/IR lowering, scheduling, memory planning, placement/routing, and executable artifacts.
Planned
Architecture
Accelerator Architecture and AI Datacenters
How are hardware resources organized?
Accelerator taxonomy, programming models, fabrics, generations, power/cooling, and design challenges.
8 sections · 23 topics
System
Training, Inference, and AI Infrastructure
How are workloads executed and operated?
Distributed training, serving, post-training, runtimes, heterogeneous fleets, and production infrastructure.
Planned
Algorithms
Models, Training Methods, and Generation Algorithms
What computation does the model require?
Model structures, training objectives, generation, compression, and model-family evolution.
Planned