GPU

Triton-Viz
Triton-Viz is an interactive visualization tool for understanding Triton programs, including execution structure, memory behavior, and GPU programming concepts in AI systems courses.
Code Demo
Linear Layouts: Robust Code Generation of Efficient Tensor Computation Using F_2
Cite Project arXiv
Triton-Sanitizer: A Fast and Device-Agnostic Memory Sanitizer for Triton with Rich Diagnostic Context
Cite Project URL
Proton: Towards Multi-level, Adaptive Profiling for Triton
Cite Project URL
Mercury: Unlocking Multi-GPU Operator Optimization for LLMs via Remote Memory Scheduling
Remote memory scheduling framework that optimizes LLM operators across multi-GPU deployments.
Cite DOI PDF
KPerfIR: Towards an Open and Compiler-centric Ecosystem for GPU Kernel Performance Tooling on Modern AI Workloads
An open, compiler-focused infrastructure for profiling and optimizing GPU kernels on AI workloads.
Cite DOI PDF arXiv
Tile-based Programming Models for AI
Introduction to Triton Programming Language for graduate students.
Apr 1, 2025 9:34 PM — 9:34 PM Cambridge, MA
Project Slides
Proton: Adaptive and Lightweight Profiling for Deep Learning Workloads
A gentle introduction to the Proton profiler and its updates.
Mar 1, 2025 9:34 PM — 9:34 PM Las Vegas, NV
Project Slides
The Proton Dialect: An MLIR Dialect For AI Compiler GPU Kernel Profiling
Presented our work on multi-level instrumentation-based profiling for Triton kernels.
Mar 1, 2025 9:34 PM — 9:34 PM Las Vegas, NV
Project Slides
Triton Update
Presented a talk about Triton and called for contributions to improving the language
Aug 13, 2024 10:56 PM — 10:56 PM Lake Tahoe, California
Project Slides