Starred repositories
Community hub for TileLang on RISC-V, bringing together compiler backends, AI kernels, developer tools, and documentation.
8
Updated Oct 26, 2026
Pure rust ML compiler with IR and support for multiple backends
openRuyi is an open source, rolling release Linux distribution for RISC-V, optimized for performance, security, AI, and cloud-native workloads.
Profiling Tools Interfaces for GPU (PTI for GPU) is a set of Getting Started Documentation and Tools Library to start performance analysis on Intel(R) Processor Graphics easily
An Open Compiler Harness for Agentic GPU Programming
Reference implementation and examples of the CuTe Layout representation and algebra.
Build compute kernels and load them from the Hub.
High-performance KV cache storage for LLM inference — GPU offloading, SSD caching, and cross-node sharing via RDMA. Works with vLLM and SGLang.
A Datacenter Scale Distributed Inference Serving Framework
Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2
DHLS (Dynamic High-Level Synthesis) compiler based on MLIR
NVIDIA's CUDA platform for Rust. Host runtime crates plus Tile (cutile-rs) and SIMT (cuda-oxide) kernel programming models in idiomatic Rust.
Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.
PyPTO kernels and end-to-end LLM models for Ascend NPUs.
[Mirror] This repository is a read-only mirror of GitCode. Please submit PRs at [https://gitcode.com/cann/pto-as](https://gitcode.com/cann/pto-as)
cuTile is a programming model for writing parallel kernels for NVIDIA GPUs
Safe, portable, high performance compute (GPGPU) kernels.
🐉 Making Rust a first-class language and ecosystem for GPU shaders 🚧
oneAPI Collective Communications Library (oneCCL)
Intel® Graphics Compute Runtime for oneAPI Level Zero and OpenCL™ Driver
The Triton backend for the PyTorch TorchScript models.