YuliangLiu0306

### What does this PR do 1. This PR construct a test to check the result of MLP with the auto parallel strategy equals to the original result or not....

### What does this PR do 1. implement search_best_logical_mesh function, which could find best logical mesh for the given device list. The best logical mesh is searched in following steps:...

### What does this PR do We implement AlphaBetaProfiler to help us detect cluster communication metrics and the best logical device mesh. In this PR, I integrate this feature into...

[autoparallel] update binary elementwise handler

1

### What does this PR do Previously, the binary elementwise handler suppose all node type argument with Tensor type meta_data, however, the type of meta_data could be int or float....

[examples] update autoparallel tutorial demo

1

This PR updates the auto-parallel tutorial demo with the lastest api.

[autoparallel] support origin activation ckpt on autoprallel system

1

### What does this PR do Previously, the communication nodes inserted by auto-parallel system will break the origin checkpoint structure, and the graphmodule generated by fx codegen will have a...

[hotfix] fix autoparallel compatibility test issues

1

## 📌 Checklist before creating the PR - [x] I have created an issue for this PR for traceability - [x] The title follows the standard format: `[doc/gemini/tensor/...]: A concise...

### 🐛 Describe the bug The compatibility test for combination auto-parallel with Gemini used same input data for different data parallel rank. Therefore, we can't check whether the gradient hooks...