Research Notes
56
Jie Liu's B-Exam: Abstractions and Optimizations for Sparse Tensor Computation on Modern Hardware
Critique of "RFSeek and Ye Shall Find: A tool for summary visualization and analysis of RFCs"
Kleene Algebra, NetKAT, StacKAT, GKAT, CF-GKAT
《经济学的思维方式》第一章读书摘录
Uncompromising Performance with Exocompilation: A Recap of Yuka Ikarashi's Talk
Paper Reading: AccelOpt: A Self-Improving LLM Agentic System for AI Accelerator Kernel Optimization
Paper Reading: Registerless Hardware Description
Paper Reading: Asynchronous Functional Reactive Programming for GUIs (The Elm Paper)
Nico Ritschel's Ph.D. Defense Summary
Pre-MICCAI Workshop@UBC Observations and Gained Insights
More...
Systems
18
Compile NEFF Executables from NKI Kernels
What is S3?
Important Locations on Jailbroken iOS
Learning MLIR and HLO by Building a Tiny StableHLO-to-LLVM IR Compiler
Using MLIR as a C++ Library with a Relocatable Install
Docker and Podman Containers as Lightweight VMs for Interactive Work
Trace Function Calls in C/C++ Programs
PRoot: Lightweight Userland Linux Distros Without Root
High-Frequency Git Operations for Everyday Development
Setting Up a Mock Windows 7 32-bit Environment Using Wine
More...
Programming
11
Rust Crates and Python Packages
Type Theory Concepts: A to Z
Personal GitHub Portfolio
Shell Coding Guidelines
C++ Coding Guidelines
Python Coding Guidelines
Metaclass Fundamentals
Creating Custom Data Structures from Abstract Containers in `collections.abc`
Type-Theoretic Constructs in C++
Python in a Functional Style: Closures, Generators, and Coroutines
More...
AI and Machine Learning
11
How the KV Cache Works in HuggingFace Transformers
Local CUDA vLLM Setup for Python-Only Development Using a Precompiled Wheel
vLLM Internals — PagedAttention and Custom Accelerator Compilation
Hugging Face Model Repositories: Organization, Semantics, and Portability
Exporting Compute Graphs, LLM Shape Dynamics, and Serving Runtimes
Schedules in Machine Learning Computation: What They Are and Who Needs to Know About Them
Main Takeaways from a Group Discussion on AI Coding
PyTorch + CUDA vs. XLA + TPU: Two Execution Models for ML Systems
Qt, OpenCV, PyTorch: The Central Dogma of GUI CV Applications
Running Local LLMs with Ollama
More...
Reflections
9
Our Cognitive Profile and a Personal Playbook for the Agentic Era
Personal Conduct
Personal Digital Governance Policy
Lessons learned from Master's thesis
Show, not tell
Some thoughts on maintaining fitness habits and improving physical fitness in the future
Strategies, Tactics, and Mindset Learned from "The Ph.D. Grind"
On Convolutional Neural Networks and Photographic Lenses
Our Motivation for Maintaining Our Blog