Since the micro architecture of AI chips are different from different vendors, so is Ascend, which is quite different from NV GPU. This repo provides the best practices of writing computing kernel using Triton Language on Ascend. To show the best practices, here provide three types of example:
- basic: basic examples
- op_extension: examples of extended operations provided by Triton Ascend
- transformer: examples used in transformer models and inference engine for LLM
- CANN
- Pytorch and Pytorch Ascend
- Triton Ascend
Follow MIT license, check LICENSE file