linuxkid473 / gbos-x86

gbOS86: Googlebook OS on ordinary Intel laptops (6th gen+): CPU-instruction emulation, patched libs, boot-time hardware detection, live/installer USB.

Panchovix / nvidia-linux-vbios-vix-loader

Load a modified VBIOS per NVIDIA GPU at boot on Linux, in memory, through the open kernel modules

0xAX / linux-insides

A book-in-progress about the Linux kernel and its insides.

tile-ai / tilelang

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

InternLM / lmdeploy

LMDeploy is a toolkit for compressing, deploying, and serving LLMs.

linkedin / Liger-Kernel

Efficient Triton Kernels for LLM Training

gmh5225 / awesome-game-security

awesome game security [Welcome to PR]

dusklinux / dusky

an unhealthy amount of work has been put into this...star it?

NVIDIA / cutile-python

cuTile is a programming model for writing parallel kernels for NVIDIA GPUs

mandiant / speakeasy

Windows kernel and user mode emulation.

deepseek-ai / TileKernels

A kernel library written in tilelang

marin-m / vmlinux-to-elf

A tool to recover a fully analyzable .ELF from a raw kernel, through extracting the kernel symbol table (kallsyms)

tenstorrent / tt-metal

:metal: TT-NN operator library, and TT-Metalium low level kernel programming model.

NVlabs / kda

Kernel Design Agents (KDA) is a agent-centric workflow to write high-performance CUDA Kernels.

ipyflow / ipyflow

A reactive Python kernel for Jupyter notebooks.

Mathics3 / mathics-core

An open-source Mathematica Kernel. This repository contains the Python modules for WL Built-in functions, variables, core primitives, e.g. Symbol, a parser to create Expressions, and an evaluator to execute them.

Dao-AILab / quack

A Quirky Assortment of CuTe Kernels

flagos-ai / FlagGems

FlagGems is an operator library for large language models implemented in the Triton Language.

CrossGL / crosstl

Translate GPU shaders and compute kernels between Metal, HLSL and GLSL, and to SPIR-V and WGSL, through the CrossGL IR

NVIDIA / cudnn-frontend

cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.

pytorch / helion

A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.

arighi / virtme-ng

Quickly build and run kernels inside a virtualized snapshot of your live system

NVIDIA / TileGym

Helpful kernel tutorials, examples and SKILLs for tile-based GPU programming

huggingface / kernels

Build compute kernels and load them from the Hub.

0xricksanchez / like-dbg

Fully dockerized Linux kernel debugging environment for Linux and MacOS

HKUSTDial / flash-sparse-attention

Trainable fast and memory-efficient sparse attention

ipython / ipykernel

IPython Kernel for Jupyter

QwenLM / FlashQLA

high-performance linear attention kernel library built on TileLang

cursor / mixture-of-kittens

Mixture-of-experts (MoE) training megakernel for NVL72s

d0u9 / Linux-Device-Driver

Advanced examples of Linux Device Drivers (LDD3) and detailed manual for running examples in QEMU which is patched with virtual PCI, USB, serial devices. I am actively composing a new book about Driver Development in Linux Kernel.

zengxiao-he / tessera

From teacher to tiles — a from-scratch LLM distillation & serving engine: custom Triton/CUDA kernels, FSDP distillation, paged-KV continuous batching, speculative decoding, a Rust gateway, a JAX oracle, and interpretability tooling.

inclusionAI / cuLA

CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.

fkodom / fft-conv-pytorch

Implementation of 1D, 2D, and 3D FFT convolutions in PyTorch. Much faster than direct convolutions for large kernel sizes.

lithos-ai / lithos-metal

lithos-metal: high-performance LLM inference kernels and serving for Apple silicon.

HezaoHezao / poirot

Poirot is a deep research agent kernel built for those who care about how agents are architected.

circle-group / hktex

Heat Kernel Textures: the Geodesic Gaussians That Do Not Splat

morluto / flameox

Runtime evidence that helps agents trace, profile, and burn down hotspots in application and native code, GPU kernels, and inference stacks.

Quchaosheng / workbench-mobile-home-robot

Evidence-first runtime for a wheeled mobile household robot with bounded actions, replayable events, and fail-closed verification.

dnv2003 / v100-skinny

Hand-written NVFP4 W4A16 CUDA kernels for Volta

Gxj230958 / TritonAscendBench

A migration benchmark that measures whether production GPU Triton kernels — extracted from vLLM and SGLang inference paths — can be ported to the Huawei Ascend 910B (Atlas A2) via `triton-ascend` without losing correctness or performance.

Inferact / tpu-megakernels

A collection of megakernels for TPU

Saganaki22 / ComfyUI-sol-attn

NVIDIA Sol-Attn for ComfyUI / Triton kernel on SM89 - SM121, with zero-copy MiniMax H3 nodes: memory-efficient attention, scheduled tau with graph preview, and feed-forward chunking. Measured 1.14–1.44× vs SageAttention and −37% MLP peak VRAM on H3

Oruk-AI / orukeet

Orukeet: multilingual ASR with fitted, frozen Gabor kernels and native inference

Blackwellboy / model-serving-minefield

Community registry of LLM serving-path traps that produce confidently wrong measurements: templates, tool parsers, reasoning fields, quant kernel paths, CUDA toolchains, KV allocation, eval harnesses, versioning. Symptom-first, with the check that catches each.

derekross / omarchy-mobile

Omarchy Mobile: native Omarchy on Google Pixel phones. Arch Linux ARM, Hyprland and Omarchy on a mainline kernel (work in progress)

tonyd2wild / DeepSeek-V4.1-Flash-vLLM-DGX-Spark

DeepSeek-V4.1-Flash (552B MoE, MXFP4 experts, 1M ctx) on four NVIDIA DGX Sparks with vLLM TP4: Engram-on-disk patch, sm121 kernel build, launchers, measured numbers

matsuzaka-yuki / susfs-lkm-module

KernelSU-compatible ZIP wrapper for pinned inforcqb SUSFS LKM; loader only, no rule management

yanght27 / GPU-Perf-Playground

GPU 性能与 AI Infra 学习项目：CUDA/Triton 算子、NCU/NSYS、vLLM/SGLang/TRT-LLM/ms-swift、PyTorch/DeepSpeed/ms-swift 训练、并行架构

MuzeAnisichael / minimal-local-agent

Small local-first agent kernel: bounded tools, reversible edits, auditable runs. CLI + Web, Ollama + compatible APIs.

Mte90 / linus-torvalds-skill

Distilled Code Reviewer skills/soul from 32k~ email from the kernel newsletter

limenagent / prodagent

A minimal, readable agent-framework kernel — understand how LangGraph/ADK work under the hood. 1800行代码的 agent 框架内核，理解 LangGraph/ADK 底层的工作原理

julianmb / halofpx

High-performance Lemonade alternative for AMD Strix Halo & Radeon — latest MoE models (Ling-3.0-Flash, Qwen 3.8 Flash Next, Ornith 1.5), bleeding-edge RDNA 3.5 Wave64/ROCmFP4 kernels, and silicon-tuned profiles.

one2piece2hello / faibench_Frontier_InfraBench

Phi-Bench (Φ-Bench): 85 open-source LLM-infrastructure engineering tasks for frontier LLMs & coding agents (KFC/LH/E2E). Self-contained public Dockerfile + offline scoring + reference solution per task. Website: llminfrabench.com

joshuaswarren / metal2vk

Compile Metal Shading Language compute kernels to Vulkan SPIR-V (clang C++ for OpenCL + a metal_stdlib shim + clspv)

Staatsgeheim / MathKernel

Evidence-aware multi-engine mathematics runtime for LLMs: exact, symbolic, formal, certified-interval and numeric computation with typed MathIR, trust labels, and full provenance

ai-infra-bench / ai-infra-bench

Evaluating frontier models on real-world ai-infra engineering tasks

arpanpathak / gpu-parallel-book

📖 GPU parallel programming from the ground up : CUDA C++ and Rust 🦀 kernels paired with clickable SVG 2D/3D architecture diagrams (warps, SMs, memory tiers). Includes reduction, scan, matmul, and a full image‑processing capstone

sudoingX / dgx-spark-deepseek-v41

DeepSeek-V4.1-Flash on two NVIDIA DGX Sparks: EXL3 weights from the official checkpoint, upstream vLLM with our patches and decode kernels

RightNow-AI / local-kimi

Optimized local serving engine for Kimi-Linear-48B: INT4 quantizer, fused decode kernels for a measured 3.18x, and an OpenAI-compatible server. Ships with k3, a bridge that detects the client per request so Claude Code, Codex, Cline, Aider and opencode all work unchanged.

Schaka / rocm-gfx803

ROCm for gfx803 aka RX 460, RX 470, RX 480, RX 550, RX 560, RX 570, RX 580 and RX 590 with custom patches, kernel fixes, etc