* refine parallel_exe
test=develop

* rename shared_var_device

* code refine

* add test_weight_decay

* remove Sort
test=develop

* Add SortForReduce
test=develop

* code refine
test=develop

* follow comment
test=develop

* refine parallel_executor

* remove uncessary code
test=develop

* wip multi process multi gpu dist training

* workable for p2p

* update test=develop

* change back env name test=develop

* fix alloc init

* fix cpu build test=devlop

* fix mac tests test=develop

* refine code

* refine test=develop

test=develop

adjust the python script

* add strategy doc
test=develop

* fix doc
test=develop

* add ParallelExecutor arg doc
test=develop

* support test using executor without reader

* run imgnet

* update fluid benchmark

* wip

* update

* update all models

* support pyreader

* update

* clean up

* make profile batches contollable

* update API.spec

* update scripts

* clean dockerfile

* update

* clean comments

* add scope argument docstring

* use num_trainers to determine nccl init comms

* hide misc APIs

* update

* fix transformer test

* update API.spec

Fix str and bytes problem
Fix sorted problem
Fix math problem
Fix CI problem

* remove block api

* remove clone_variable

* hide block inner apis

* update

* fix tests

* change return_numpy[PE] default value

* Remove convert to numpy in unit test

replace use_event with use_cuda, because use_event means the program running with CUDA, so use_cuda maybe more intuitive.