Kernelize is Powered by Triton

Triton is the key enabling technology that makes day-0 support for new models possible. Kernelize uses Triton to generate optimized kernels for new hardware targets, enabling vLLM and Ollama to run on new hardware alternatives.

Day-0 Support Enables Lower Costs

By using Triton to generate optimized kernels for new hardware devices, Kernelize provides day-0 support for new models at a fraction of the cost.

What is Triton?

Triton is a language and compiler for parallel programming that aims to provide a Python-based programming environment for productively writing custom DNN compute kernels capable of running at maximal throughput on modern GPU hardware. It's the enabling technology that makes day-0 support for new models possible across diverse hardware.

Python-Based Programming

Write custom DNN compute kernels in Python with a familiar programming environment.

Maximal Throughput

Achieve optimal performance on modern GPU hardware with compiler optimizations.

Parallel Programming

Language and compiler designed specifically for parallel computing workloads.

Open Source

MIT licensed with an active community of contributors and users.