Highlights
- Pro
Stars
Open ML Compiler Foundation for the Agentic Era
UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g., GPU-driven)
NVIDIA's CUDA platform for Rust. Host runtime crates plus Tile (cutile-rs) and SIMT (cuda-oxide) kernel programming models in idiomatic Rust.
DeepInfer: High-performance, large-scale inference framework for DeepSeek models
Automatic formal verification for warp-specialized GPU kernels.
Lightweight Benchmark Server for Agentic GPU Programming
An Open Compiler Harness for Agentic GPU Programming
Kernel Design Agents (KDA) is a agent-centric workflow to write high-performance CUDA Kernels.
A tutorial on modern GPU programming for machine learning systems
Fast and memory-efficient classical machine learning operators
A kernel library written in tilelang
Our first fully AI generated deep learning system
A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.
Fast and memory-efficient exact attention
cuTile is a programming model for writing parallel kernels for NVIDIA GPUs
Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.
Perplexity open source garden for inference technology
Building the Virtuous Cycle for AI-driven LLM Systems
Ship correct and fast LLM kernels to PyTorch
A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.
An extremely fast Python package and project manager, written in Rust.