halide/Halide

a language for fast, portable data-parallel computation

What it solves

Halide is designed to simplify the creation of high-performance image and array processing code. It removes the complexity of manually optimizing for modern hardware architectures, allowing developers to write efficient code that runs across a wide variety of platforms.

How it works

Halide is an embedded language (available in C++ and Python) rather than a standalone one. Developers use the Halide API to build an in-memory representation of a processing pipeline. This pipeline can then be compiled into an object file (Ahead-of-Time compilation) or JIT-compiled and executed within the same process.

Who it’s for

It is intended for developers working on image and array processing who need high-performance execution on diverse hardware, including CPUs (X86, ARM, RISC-V, etc.) and GPUs (CUDA, OpenCL, Metal, Vulkan, etc.).

Highlights

  • Broad Hardware Support: Targets a vast array of CPU architectures and GPU Compute APIs.
  • C++ and Python Bindings: Offers full support for writing pipelines in either language.
  • Flexible Compilation: Supports both JIT and AOT compilation modes.
  • Cross-Platform: Compatible with multiple operating systems including Linux, Windows, macOS, Android, and iOS.

Related

  • Project
    ROCm/hip

    HIP is a C++ runtime API and kernel language for AMD GPUs that enables developers to port NVIDIA CUDA code to AMD hardware for portable GPU acceleration.

  • Project
    ermig1979/Simd

    A high-performance C/C++ image processing and machine learning library that uses SIMD CPU extensions to accelerate algorithms across x86, ARM, and Hexagon architectures.

  • Project
    facebookincubator/AITemplate

    A Python framework that transforms deep neural networks into optimized CUDA/HIP C++ code for high-performance inference serving on NVIDIA and AMD GPUs.

  • Project
    CVCUDA/CV-CUDA

    A GPU-accelerated library of computer vision algorithms designed to provide high-throughput, low-latency image and video processing for AI pipelines on NVIDIA platforms.

  • Project
    google/highway

    A C++ library providing portable SIMD/vector intrinsics to enable high-performance, architecture-independent software across various CPU platforms.