The examples in here showcase both the GPU side and the CPU side of writing a tool which uses the GPU.
A simple vector addition example demonstrating basic CUDA kernel usage.
General Matrix Multiplication (GEMM) implementation showing more complex CUDA operations.
Demonstrates using an existing Rust crate (sha2 from crates.io) on both CPU and GPU without modification. Shows that the same cryptographic hashing code can run on CUDA, producing identical results to the CPU implementation.
A very simple interactive Path Tracer inspired by Ray Tracing In One Weekend which runs on CPU or GPU, with the additional option of running OptiX denoising.
The Path Tracer uses cuda_builder to compile the core path tracer for the GPU and GPU (hardware raytracing), and uses the core path tracer as a normal crate for CPU rendering and sharing structures.
Example of computes the dot product of two single-precision (f32) vectors