RenderKit/embree

Embree ray tracing kernels repository.

What it solves

Intel® Embree is a high-performance ray tracing library designed to improve the performance of photo-realistic rendering applications. It specifically addresses the challenges of production rendering, focusing on incoherent ray performance, high-quality acceleration structure construction, and low memory consumption.

How it works

Embree provides optimized ray tracing kernels for x86 CPUs (supporting SSE, AVX, AVX2, and AVX-512) and Intel GPUs (via the SYCL open standard). It uses runtime code selection to choose the best kernel for the hardware. The library supports a wide range of primitive types, including triangles, Catmull-Clark subdivision surfaces, and various curve primitives. It also implements high-performance two-level spatial index structure construction algorithms to support dynamic scenes.

Who it’s for

Graphics application developers who are building photo-realistic renderers and need a high-performance, hardware-accelerated ray tracing backend.

Highlights

  • Broad Hardware Support: Supports x86 CPUs (Linux, macOS, Windows), ARM CPUs (Linux, macOS), and Intel GPUs (Linux, Windows).
  • Rich Primitive Set: Includes support for triangles, subdivision surfaces, curves, and user-defined procedural geometries.
  • Advanced Rendering Features: Supports motion blur (multi-segment, deformation, and quaternion), ray masking, and multi-level instancing.
  • Flexible Integration: Provides an ISPC interface for automatic vectorization and SYCL support for a single source codebase that runs on both CPUs and GPUs.

Related

  • Dispatch
    CPU Optimized Embeddings with Optimum Intel and fastRAG

    Hugging Face and Intel have introduced a method to accelerate embedding models on Xeon CPUs using Optimum Intel and fastRAG, achieving up to 4.5x latency reduction and 4x throughput improvement via int8 quantization.

  • 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.

  • Dispatch
    Intel and Hugging Face Partnership for Machine Learning Hardware Acceleration

    Intel has joined the Hugging Face Hardware Partner Program to develop the Optimum Intel library, enabling streamlined hardware acceleration for training, fine-tuning, and inference of Transformer models on Intel platforms.

  • Project
    OpenMathLib/OpenBLAS

    An optimized Basic Linear Algebra Subprograms (BLAS) library that provides high-performance linear algebra routines across a wide variety of CPU architectures.

  • Project
    google/XNNPACK

    XNNPACK is a highly optimized library of low-level performance primitives used to accelerate neural network inference across ARM, x86, WebAssembly, and RISC-V platforms.