* change Vector to std::vector and provide MixVector class as a helper wrapper class

* solve the multi-gpu hang problem

* remove the duplicate template instantialize

* Copy vector to cpu

* add CopyToCPU

* xxx

* final version: fix the problem of all reduce

* remove mixvector dependence

* fix

* merge

* fix code

* fix by CI

* Replaced pten::LoD with paddle::framework::LoD

* Overrided CPUVector with CUDAVector

* Refactored paddle::framework::Vector