LLM training in simple, raw C/CUDA

Instant neural graphics primitives: lightning fast NeRF and more

A massively parallel, optimal functional runtime in Rust

DeepEP: an efficient expert-parallel communication library

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

CUDA accelerated rasterization of gaussian splatting

Tile primitives for speedy kernels

[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

[ARCHIVED] Cooperative primitives for CUDA C++. See https://github.com/NVIDIA/cccl

[MICRO'23, MLSys'22] TorchSparse: Efficient Training and Inference Framework for Sparse Convolution on GPUs.

FSA/FST algorithms, differentiable, with PyTorch compatibility.

[CVPR 2023 Highlight] Neural Kernel Surface Reconstruction

[ICML2025] SpargeAttention: A training-free sparse attention that accelerates any model inference.

[ICCV 2023] Official code for NeuS2

UNet diffusion model in pure CUDA

Causal depthwise conv1d in CUDA, with a PyTorch interface

Original implementation of "Radiant Foam: Real-Time Differentiable Ray Tracing"

Instant neural graphics primitives: lightning fast NeRF and more

State of the art sorting and segmented sorting, including OneSweep. Implemented in CUDA, D3D12, and Unity style compute shaders. Theoretically portable to all wave/warp/subgroup sizes.

Differentiable Iso-Surface Extraction Package (DISO)

Differentiable gaussian rasterization with depth, alpha, normal map and extra per-Gaussian attributes, also support camera pose gradient

3DGS-LM accelerates Gaussian-Splatting optimization by replacing the ADAM optimizer with Levenberg-Marquardt. (ICCV 2025)

Marching cubes implementation for PyTorch environment.

A modular differential gaussian rasterization library.

[ECCV'24] On the Error Analysis of 3D Gaussian Splatting and an Optimal Projection Strategy

A differentiable rasterizer used in the project "2D Gaussian Splatting"

Official code release for "Efficient Perspective-Correct 3D Gaussian Splatting Using Hybrid Transparency"