
Guides an agent through iterative profiling and tuning of cuTile GPU kernels in TileGym
- Open source
- 4K GitHub stars
- Developer tools


Guides an agent through iterative profiling and tuning of cuTile GPU kernels in TileGym

Guides an agent to write, debug and validate cuTile GPU kernels in Python

Guides an agent in converting cuTile GPU kernels to Triton and debugging the result

Guides an agent in porting cuTile Python GPU kernels to Julia cuTile.jl

Guides an agent through adding a new cuTile GPU kernel operator to TileGym

Guides an agent to swap TileGym cuTile kernels into Hugging Face transformers models

Guides an agent to find CUDA memory, shared-memory race and sync bugs with Compute

Guides an agent through GPU DataFrame work with NVIDIA cuDF and dask-cuDF

Guides setup, GPU simulation targets, QPU access and @cudaq.kernel authoring in CUDA-Q

Guides an agent through building a first single-GPU cuDF path in a query engine

Guides an agent through modifying, building, testing and contributing to the cuOpt

Guides installing NVIDIA cuOpt for Python, C, or REST server use via pip, conda or Docker
The closest alternatives to Tilegym cutile autotuning are Tilegym improve cutile kernel perf, Tilegym cutile python, Tilegym converting cutile to triton, Tilegym converting cutile to julia and Tilegym adding cutile kernel.
Yes. Tilegym improve cutile kernel perf, Tilegym cutile python, Tilegym converting cutile to triton have a free or open-source option.