Schedule

Computing for AI AI for Computing

13 Friday meetings, 10:10 AM - 12:00 PM, MUDD Building 825. Presentation slides are due 11:59 PM the Thursday before class.

01
Launch

Course Introduction: AI as Workload, AI as Designer

Lecture slides (PDF)

Research question of the weekHow does the relationship between workloads and architectures change when AI can adapt both?

Instructor-selected background · no student presentation

  1. A New Golden Age for Computer Architecture CACM 2019
  2. GenAI for Systems: Recurring Challenges and Design Principles from Software to Silicon arXiv 2026
Optional readings (2)

Module 1: Computing for AI

02
Computing for AI

How to Study an AI Computing System + LLM Inference Fundamentals

Lecture slides (PDF)

Research question of the weekWhat makes a performance model useful when bottlenecks shift with workload, scale, and operating conditions?

Instructor-selected background · no student presentation

  1. Roofline: An Insightful Visual Performance Model for Multicore Architectures CACM 2009
  2. The Tail at Scale CACM 2013
  3. FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness NeurIPS 2022
Optional readings (2)
03
Computing for AI🎤 Guest Speaker

Serving Systems for LLMs and AI Agents

Lecture slides (PDF)

Guest lecture: Prof. Jovan Stojkovic (UT Austin)

Research question of the weekWhat is the right unit of optimization when an AI application is a workflow rather than a single model?

Mini-lecture case study: DyServe - a workflow-aware serving layer (instructor's work)

Student-led papers

  1. Efficient Memory Management for Large Language Model Serving with PagedAttention SOSP 2023
    KV-cache virtualization, fragmentation, batching, vLLM

    Presenter: Aaron Cherian · Presentation slides (PPTX)

  2. SGLang: Efficient Execution of Structured Language Model Programs NeurIPS 2024
    RadixAttention, KV reuse across program structure

    Presenters: Gavin Zou, Pingchuan Dong · Presentation slides (PPTX)

Optional readings (12)
  • Project teams formed
04
Computing for AI🎤 Guest Speaker

AI Hardware: Accelerators and System-on-Chips

Lecture slides (PDF)

Guest lecture: Prof. Tushar Krishna (Georgia Tech)

Research question of the weekWhen does specialization create a lasting advantage, and when does it bind a system too closely to today's workloads?

Student-led papers

  1. In-Datacenter Performance Analysis of a Tensor Processing Unit ISCA 2017
    read as a ten-year retrospective

    Presenters: Lizhong Wang, Jiangrui Xu · Presentation slides (PPTX)

  2. Open AI Jalapeño Chip Hot Chips 2026 slides OpenAI blog
    inference-driven ASIC and system co-design

    Presenters: Shen Li, Yichen Xu · Presentation slides (PPTX)

Optional readings (8)
05
Computing for AI🎤 Guest Speaker

Physical AI I: Inference Systems and Serving for Embodied AI

Lecture slides (PDF)

Guest lecture: Prof. Yilun Du (Harvard)

Research question of the weekHow should computation be organized when the value of a prediction depends on when an action is taken?

Student-led papers

  1. Embodied.cpp: A Portable Inference Runtime of Embodied AI Models on Heterogeneous Robots arXiv 2026
    deployable embodied-AI runtime across heterogeneous robots

    Presenters: Ryan Ma, George Wang · Presentation slides (PPTX)

  2. How Fast Can I Run My VLA? Demystifying VLA Inference Performance with VLA-Perf arXiv 2026
    analytical performance model; device/edge/cloud placement

    Presenters: Shiyao Marcus Lam, Pulak Mehrotra · Presentation slides (PDF)

Optional readings (7)
06
Computing for AI🎤 Guest Speaker

Physical AI II: Hardware-Software Co-Design and Architecture

Guest lecture: Prof. Luca Carloni (Columbia)

Research question of the weekWhere should the boundary between algorithm and hardware move when latency, energy, and task success must be optimized together?

Student-led papers

  1. ReCA: Integrated Acceleration for Real-Time and Efficient Cooperative Embodied Autonomous Agents ASPLOS 2025
    cross-layer co-design for cooperative embodied agents
  2. DiTPA: A DiT-Based Action Planner Accelerator Exploiting Action-Denoising-Multimodality Redundancy for Embodied Artificial Intelligence ISCA 2026
    exploiting action, denoising, and multimodality redundancy for embodied-AI action planning
Optional readings (8)
  • Oct 19, 11:59 PM - P1 infrastructure & baselines
07
Computing for AI

Accelerating Neuro-Symbolic and Compositional AI

Research question of the weekHow should computing abstractions change when intelligence combines learning, search, and symbolic reasoning?

Student-led papers

  1. CogSys: Efficient and Scalable Neurosymbolic Cognition System via Algorithm-Hardware Co-Design HPCA 2025
    workload characterization, reconfigurable processing
  2. REASON: Accelerating Probabilistic Logical Reasoning for Scalable Neuro-Symbolic Intelligence HPCA 2026
    unified DAGs, irregular reasoning acceleration
  3. Lobster: A GPU-Accelerated Framework for Neurosymbolic Programming ASPLOS 2026
    compiling Datalog-based neurosymbolic programs to GPUs
Optional readings (4)

Module 2: AI for Computing

08
AI for Computing

AI for Software Systems: Compilers and GPU Kernels

Research question of the weekWhat should an AI optimizer be free to change, and which properties must remain invariant?

Student-led papers

  1. Faster Sorting Algorithms Discovered Using Deep Reinforcement Learning (AlphaDev) Nature 2023
    algorithm discovery, reward design, validation
  2. AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery DeepMind 2025
    evolutionary LLM code search across scheduling, TPU circuits, kernels
  3. KernelBench: Can LLMs Write Efficient GPU Kernels? ICML 2025
    benchmark design; discussed with the Sakana AI CUDA Engineer episode
Optional readings (5)
  • Nov 2, 11:59 PM - P2 prototype & pilot results
09
Project

Midterm Project Presentations

Research question of the weekWhich uncertainty is most worth resolving before investing further in a research direction?

Team presentations; no mini-lecture and no assigned papers.

  • Nov 5, 11:59 PM - midterm slides
10
AI for Computing🎤 Guest Speaker

AI for Computer Architecture I: Measuring and Exploring

Guest lecture: Prof. Sagar Karandikar (UC Berkeley)

Research question of the weekWhat makes a design environment informative enough for an AI agent to discover better architectures?

Student-led papers

  1. ArchEval: Measuring AI Agents as Computer Architects arXiv 2026
    benchmarking LLM agents on architecture tasks
  2. CHIA: An Open-Source Framework for Principled, Agentic AI-Driven Hardware/Software Co-Design Research arXiv 2026
    composable design loops for agent-driven HW/SW co-design
Optional readings (8)
11
AI for Computing🎤 Guest Speaker

AI for Computer Architecture II: Architecture 2.0

Guest lecture: Prof. Vijay Janapa Reddi (Harvard)

Research question of the weekWhat would it take for AI-native design to turn high-level intent into better, physically verified systems and chips?

Student-led papers

  1. ArchOrchestra: End-to-End Agentic Accelerator Design Forthcoming 2026
    end-to-end agentic accelerator design
  2. Architecture 2.0: Principles of AI-Native System and Chip Design Online book, 2026
    Read one chapter.
Optional readings (7)
  • Nov 23, 11:59 PM - P3 evaluation plan & initial results

No class - Thanksgiving

12
AI for Computing

AI for RTL and Chip Physical Design

Research question of the weekHow should exploration and verification work together when AI makes consequential design decisions?

In-class exercise: In-class critical reading: the RedWood white paper (Architect Labs 2026) audited with the evidence-centered checklist

Student-led papers

  1. A Graph Placement Methodology for Fast Chip Design (AlphaChip) Nature 2021 (+2024 addendum) co-led by a pair
    reinforcement learning for chip placement; co-led by a pair
  2. ChipNeMo: Domain-Adapted LLMs for Chip Design NVIDIA 2024
    domain adaptation, retrieval, EDA scripts
  3. Comprehensive Verilog Design Problems: A Next-Generation Benchmark Dataset for Evaluating Large Language Models and Agents on RTL Design and Verification arXiv 2025
    RTL benchmarks and their limitations
Optional readings (9)
  • Dec 7, 11:59 PM - P4 main results & ablations
13
Synthesis

Final Project Poster Session

Research question of the weekWhat general principle does a research result establish, and where does that principle stop applying?

Poster showcase; no mini-lecture and no assigned papers.

  • Dec 10, 11:59 PM - posters in shared folder
  • Dec 11, 11:59 PM - P5 complete draft & artifact
  • Dec 18, 11:59 PM - final paper & artifact

Registrar dates: Sep 18: Last day to add (Change of Program) · Oct 13: Drop deadline · Nov 19: Pass/fail & withdrawal deadline · Nov 27: No class - Thanksgiving · academic calendar