LLM training in simple, raw C/CUDA

A massively parallel, optimal functional runtime in Rust

RAFT contains fundamental widely-used algorithms and primitives for machine learning and information retrieval. The algorithms are CUDA-accelerated and form building blocks for more easily writing …

Distributed multigrid linear solver library on GPU

Nonuniform fast Fourier transforms of types 1 and 2, in 1D, 2D, and 3D, on the GPU