HanHaowen Blog

Thinking will not overcome fear but action will.

The Principle of Operator Registration in PyTorch

Note: This is a beginner’s article, feedback is welcome! The following content is based on PyTorch 2.0.0. In the official PyTorch tutorial Extend Dispatcher, the main way to register operators is ...

Posted by HanHaowen on December 21, 2023

Selection of block size and grid size, as well as occupancy calculation, in CUDA programming.

Note: Beginner’s article, feedback is welcome. Chinese version: https://juejin.cn/post/7312338839695114267 In CUDA programming, occupancy refers to the ratio of active thread warps to the maximum...

Posted by HanHaowen on December 16, 2023

FEATURED TAGS
CUDA Artificial Intelligence C++
ABOUT ME

Both my bachelor's and master's degrees were obtained from Harbin Institute of Technology, with a major in Optical Engineering and Electronics. Currently, I am working at Moore Threads as a developer focusing on AI infrastructure development.

✉️ haowen.han@mthreads.com