KernelTuner/kernel_tuner

Kernel Tuner

65
/ 100
Established

Optimize your GPU applications to run faster and more efficiently. You provide your existing GPU kernel code in languages like CUDA, HIP, or OpenCL, and this tool systematically tests different configurations to find the best-performing version. This is ideal for GPU programmers and performance engineers looking to maximize the speed and energy efficiency of their high-performance computing tasks.

389 stars. Available on PyPI.

Use this if you need to automatically fine-tune the performance of individual GPU code snippets (kernels) to achieve faster execution or lower energy consumption.

Not ideal if you are looking for a high-level GPU programming abstraction and do not want to work directly with low-level kernel code.

GPU programming high-performance computing CUDA optimization OpenCL tuning energy efficiency
Maintenance 10 / 25
Adoption 10 / 25
Maturity 25 / 25
Community 20 / 25

How are scores calculated?

Stars

389

Forks

63

Language

Python

License

Apache-2.0

Last pushed

Mar 12, 2026

Commits (30d)

0

Dependencies

8

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/KernelTuner/kernel_tuner"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.