The basic idea is: exploits data dependency to control the execution order
of side-effect operations, and keep the semantics of ANF unchanged.

The ControlDepend primitive is removed and there are two primitives added:

1. UpdateState:
```
  a = Assign(para, value)
```
became:
```
  a = Assign(para, value, u)
  u = UpdateState(u, a)
```

2. Load:
```
  x = Add(para, value)
```
became:
```
  p = Load(para, u)
  x = Add(p, value)
  u = UpdateState(u, p)
```

From: @lijiaqi0612
Reviewed-by: @c_34,@zh_qh
Signed-off-by: @c_34

From: @yao_yf
Reviewed-by: 
Signed-off-by:

From: @Somnus2020
Reviewed-by: @kingxian
Signed-off-by: @kingxian

Merge pull request !6259 from lijiaqi/add_uttest

Merge pull request !6144 from lijiaqi/add_validation

Merge pull request !6101 from wangnan39/delete_datawrapper

Merge pull request !5063 from lijiaqi/momentum_and_sgd

Merge pull request !4121 from Simson/enhancement-API