m4rs-mt/ILGPU

ILGPU JIT Compiler for high-performance .Net GPU programs

What it solves

ILGPU is a JIT (just-in-time) compiler that allows developers to write high-performance GPU programs using .NET-based languages (like C#) instead of relying on native dependencies or specialized GPU languages like CUDA C++.

How it works

It compiles .NET code into GPU-executable instructions. It provides a multi-threaded CPU accelerator, which allows kernels to be executed and debugged on the CPU before being deployed to the GPU. It also includes a standard library of high-level algorithms (such as sorting and prefix sum) that work across all supported accelerator types.

Who it’s for

.NET developers who need the performance of GPU acceleration for their applications without leaving the C# ecosystem or needing to write native C++ code.

Highlights

  • Pure C# Implementation: Written entirely in C# with no native dependencies.
  • CPU Debugging: Integrated CPU accelerator for debugging kernels and hardware features like shared memory and atomics.
  • C# Integration: Kernels can use default C# functions and value types without special annotations.
  • Algorithm Library: Includes a built-in library of high-level algorithms for common GPU tasks.

Related

  • Project
    mitsuba-renderer/drjit

    A just-in-time compiler for ordinary and differentiable computation designed to handle the massive computation graphs typical of differentiable rendering.

  • Project
    brainpy/BrainPy

    A JIT-compiled framework for computational neuroscience and brain-inspired computation that enables efficient simulation, training, and analysis of brain dynamics models.

  • Project
    Jittor/jittor

    Jittor is a high‑performance deep‑learning framework that JIT‑compiles the entire computation graph and uses meta‑operators to generate specialised CUDA/C++ kernels. It offers a PyTorch‑like Python API, supports CPU and GPU back‑ends (CUDA, ROCm, Hygon), and includes a model zoo for vision, rendering, geometric learning, and reinforcement learning.

  • Project
    gpujs/gpu.js

    A JavaScript acceleration library that transpiles JS functions into shader language to enable GPGPU (General Purpose computing on GPUs) in the browser and Node.js.

  • Project
    tracel-ai/cubecl

    A Rust language extension and JIT compiler for writing high-performance compute kernels that run across NVIDIA, AMD, Apple, and CPU platforms.