Crate cuda_rust_wasm

CUDA-Rust-WASM: A CUDA to Rust transpiler with WebGPU/WASM support

This crate provides a complete toolchain for translating CUDA code to Rust, with support for WebGPU backends and WASM compilation.

Re-exports

pub use error::CudaRustError;

pub use error::Result;

pub use parser::CudaParser;

pub use transpiler::Transpiler;

pub use transpiler::CudaTranspiler;

pub use runtime::Runtime;

pub use neural_integration::NeuralBridge;

pub use neural_integration::BridgeConfig;

pub use neural_integration::NeuralOperation;

pub use neural_integration::ActivationFunction as NeuralActivationFunction;

pub use neural_integration::SystemCapabilities as NeuralCapabilities;

pub use neural_integration::initialize as init_neural_integration;

pub use neural_integration::get_capabilities as get_neural_capabilities;

Modules

backend

Backend abstraction layer

error

Error types for CUDA-Rust transpiler

kernel

Kernel execution module

memory

Memory management module

neural_integration

Neural Network Integration for CUDA-WASM with ruv-FANN

parser

CUDA code parsing module

prelude

Prelude module for convenient imports

profiling

Performance profiling tools for cuda-rust-wasm

runtime

CUDA-compatible runtime for Rust

transpiler

CUDA to Rust transpilation module

utils

Utility functions and helpers

Macros

kernel_function

Helper macro to define kernel functions

memory_error

Macro for creating memory errors

parse_error

Helper macro for creating parse errors

profile_scope

Macro for easy profiling

runtime_error

Helper macro for creating runtime errors

time_block

Macro for easy timing of code blocks

translation_error

Helper macro for creating translation errors

Structs

CudaRust

Main entry point for transpiling CUDA code to Rust

Functions

init

Initialize the CUDA-Rust runtime