2025年阅读组会

Redundancy-Aware Decoding and Inference:Minimizing Video Analytics Overheads

ICRA'2024

DeepSeek-Coder: When the Large Language Model Meets

Programming - The Rise of Code Intelligence

Optimistic Stack Allocation and Dynamic 

Heapification for Managed Runtimes

PLDI'24

Falcon: A Fused Approach to Path-Sensitive 

Sparse Data Dependence Analysis

Apollo: An Exploration of Video Understanding in 

Large Multimodal Models

External Large Foundation Model: How to

Efficiently Serve Trillions of Parameters for Online

Ads Recommendation

On-Demand Scenario Generation for Testing Automated Driving Systems

The AI CUDA Engineer: Agentic CUDA Kernel

Discovery, Optimization and Composition

Flex Attention: A Programming Model for 

Generating Optimized Attention Kernels

with Multiversion Dependency Graphs

FCoT-VL:Advancing Text-oriented Large Vision-Language 

Models with Efficient Visual Token Compression

Using Graph Databases to Investigate Trends in 

Structure−Activity Relationship Networks

ISCA'25

FlashInfer: Efficient and Customizable Attention 

Engine for LLM Inference Serving

Datalog-Based Language-Agnostic Change Impact  

Analysis for Microservices

Optimization for 2D Trapped-Ion Quantum

Look Ma, No Bubbles! Designing a Low-Latency 

Megakernel for Llama-1B

CAD-Llama: Leveraging Large Language Models for 

Computer-Aided Design  Parametric 

3D Model Generation

HPCTransCompile: An AI Compiler Generated 

Dataset for High-Performance CUDA Transpilation 

and LLM Preliminary Exploratio