| 1 | Aug. 25 (Tue) | Course Overview and Logistics | ShiftAddLLM: Accelerating Pretrained LLMs via Post-Training Multiplication-Less Reparameterization, NeurIPS'24
EyeCoD: Eye Tracking System Acceleration via FlatCam-based Algorithm & Accelerator Co-Design, ISCA'22
Drawing Early-Bird Tickets: Towards More Efficient Training of Deep Networks, ICLR'20
HiLo-Token: Input-Adaptive High-Low Frequency Token Compression for Efficient Image Editing Model Serving, arXiv'26 |
| Aug. 27 (Thu) | Intro to ML System Co-Design |
| 2 | Sep. 1 (Tue) | Paper Discussion | Kairos: A Scalable Serving System for Physical AI, arXiv'26 |
| Sep. 3 (Thu) | Invited Talk | MixFusion: A Patch-Level Parallel Serving System for Mixed-Resolution Diffusion Models, PPoPP'26
FlexCache: Flexible Approximate Cache System for Video Diffusion, arXiv'24
Attacks on Approximate Caches in Text-to-Image Diffusion Models, Usenix Sec’26 |
| 3 | Sep. 8 (Tue) | Paper Discussion | Towards High-Goodput LLM Serving with Prefill-decode Multiplexing, ASPLOS'26 |
| Sep. 10 (Thu) | Paper Discussion | DistServe: Disaggregating Prefill and Decoding for Goodput-optimized Large Language Model Serving, OSDI'24 |
| 4 | Sep. 15 (Tue) | Paper Discussion | Patterns behind Chaos: Forecasting Data Movement for Efficient Large-Scale MoE LLM Inference, ISCA'26 |
| Sep. 17 (Thu) | Paper Discussion | Cassandra: Enabling Reasoning LLMs at Edge via Self-Speculative Decoding, ISCA'26 |
| 5 | Sep. 22 (Tue) | Paper Discussion | Combating the Memory Walls: Optimization Pathways for Long-Context Agentic LLM Inference, ISCA'26 |
| Sep. 24 (Thu) | Paper Discussion | Tetris: Efficient Long-context LLM Serving with Chunkwise Dynamic Sequence Parallelism, ISCA'26 |
| 6 | Sep. 29 (Tue) | Paper Discussion | ExecuTorch -- A Unified PyTorch Solution to Run AI Models On-Device, MLSys'26 |
| Oct. 1 (Thu) | Paper Discussion | StreamDiffusionV2: A Streaming System for Dynamic and Interactive Video Generation, MLSys'26 |
| 7 | Oct. 6 (Tue) | Paper Discussion | Aegaeon: Effective GPU Pooling for Concurrent LLM Serving on the Market, SOSP'25 |
| Oct. 8 (Thu) | Paper Discussion | Kaleido: Algorithm-Hardware Co-Design for Video DiTs by Exploiting Latent Space Correlations, MICRO'26 |
| 8 | Oct. 13 (Tue) | Fall Break |
| Oct. 15 (Thu) | Paper Discussion | HeteroMosaic: Exposing and Exploiting Heterogeneous Opportunities for Energy-Efficient Edge LLM Inference, MICRO'26 |
| 9 | Oct. 20 (Tue) | Paper Discussion | Beyond Utilization: Energy-Conscious GPU Sharing for Inference Serving, SOSP'26 |
| Oct. 22 (Thu) | Paper Discussion | DiFlow: A System for Micro-Serving Text-to-Image Diffusion Workflows, SOSP'26 |
| 10 | Oct. 27 (Tue) | Paper Discussion | Sharpen the Spec, Cut the Code: A Case for Generative File System with SYSSPEC, FAST'26 |
| Oct. 29 (Thu) | Paper Discussion | KerColle: Unlocking Fine-Grained GPU Concurrency in Vision-Language-Action Models, PACT'26 |
| 11 | Nov. 3 (Tue) | Paper Discussion | Xtrace: High-Fidelity GPU Intra-Kernel Tracing via Binary-Level Instruction Splicing, arXiv'26 |
| Nov. 5 (Thu) | Invited Talk | AVO: Agentic Variation Operators for Autonomous Evolutionary Search, arXiv'26
AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery, arXiv'25 |
| 12 | Nov. 10 (Tue) | Paper Discussion | SPLASH: Co-Designing Sparse Attention with High-Bandwidth Flash for Efficient Long-Context Inference, arXiv'26 |
| Nov. 12 (Thu) | Paper Discussion | Optimus: Elastic Decoding for Efficient Diffusion LLM Serving, arXiv'26 |
| 13 | Nov. 17 (Tue) | Paper Discussion | ParallelKittens: Systematic and Practical Simplification of Multi-GPU AI Kernels, MLSys'26 |
| Nov. 19 (Thu) | Paper Discussion | CAKE: Compiler-Agent Co-Design for Frontier Kernel Evolution, arXiv'26 |
| 14 | Nov. 24 (Tue) | Paper Discussion | db-SP: Accelerating Sparse Attention for Visual Generative Models with Dual-Balanced Sequence Parallelism, MLSys'26 |
| Nov. 26 (Thu) | Thanksgiving Break |
| 15 | Dec. 1 (Tue) | Final Project Presentation |
| Dec. 3 (Thu) |
| 16 | Dec. 8 (Tue) |
| Dec. 10 (Thu) |
| 17 | Dec. 15 (Tue) | Exam Week (No Schedule) |
| Dec. 17 (Thu) |