EmbarkStudios/rust-gpuArchived

Features

• SPIR-V Translators - Translates Rust intermediate representation into SPIR-V bytecode using a custom LLVM-based compiler backend.
• Compute Shader Programming - Supports writing general-purpose GPU compute kernels in Rust for parallel data processing workloads.
• Cross-Platform Shader Compilers - Provides a compilation target that enables Rust code to run across different GPU hardware via SPIR-V.
• GPU Compute Toolchains - Offers a complete development environment for writing parallel compute kernels and graphics shaders in Rust.
• GPU Memory Layout Mappings - Translates high-level Rust types into hardware-compatible memory layouts and GPU-specific address spaces.
• LLVM-Based Compilers - Implements a custom compiler backend using LLVM to map Rust IR to GPU-compatible memory layouts.
• Multi-Target Compilers - Implements a compiler target model that translates Rust source into bytecode for non-CPU GPU execution environments.
• GPU Intrinsics - Exposes GPU-specific operations such as barriers and texture sampling as native Rust intrinsic functions.
• Shader Intrinsic Mappings - Maps GPU-specific hardware operations like texture sampling and barriers to native Rust intrinsic functions.
• Real-Time Graphics Development - Enables the development of vertex, fragment, and other graphics shaders in Rust for real-time rendering.
• Shader Entry Point Configurators - Provides mechanisms to identify and configure specific Rust functions as entry points for various shader stages.
• Zero-Cost Abstraction Patterns - Compiles high-level Rust loops and branches directly into SPIR-V instructions without runtime overhead.

Star history