Pinned Loading

1. gpgpu-arena gpgpu-arena Public

   Benchmark and compare CUDA kernel implementations side-by-side.

   C++

2. microgpt.cpp microgpt.cpp Public

   A minimal GPT implementation in C++ with scalar autograd that runs fast on a single thread.

   C++ 25 3

3. sim.warp_nbody sim.warp_nbody Public

   An omniverse extension for simulating and rendering the N-body problem using NVIDIA’s WARP GPU DSL

   Python

4. cuopt.piperouter cuopt.piperouter Public

   Omniverse extension that routes pipes through 3D environments using NVIDIA Warp for GPU voxelization and cuOpt for route optimization.

   Python

5. cutile-flashattn cutile-flashattn Public

   Flash attention forward kernel in NVIDIA CuTile

   Python

6. gpu-anatomy gpu-anatomy Public

   Cuda