* 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

simplify analysis tests

fix test issues on windows

Accelerate PADDLE_ENFORCE

Fix conv_elementwise_add2_act pass

Revert "cherry-pick the #12759"

fix ci error. test=develop

Add scope_pool

Add seq pool1 inference test

test=develop

test=develop

test=develop

add module cleanup
test=develop

fix bug in if-else op

test=develop

test=develop

This reverts commit 7f6d8ace.

test=develop

* code style fix, test=develop

test=develop, fix ci by install requirement and add pip install valid…

test=develop

Fix/abnormal stack op time

* Make topk op support variable k.
test=develop

* Fix tensor type.
test=develop

cherry-pick the #12759

* wip

* wip

* wip

* wip for test

* add fp16 tests test=develop

* fix cpu build test=develop

* fix test=develop

* fix py3 tests test=develop

* fix lr_scheduler dtype test=develop

* fix test=dvelop

* test fix ci compile test=develop

* fix build and merge test=develop

* fallback momentumop change to general test=develop

* make fp16 lr schedule simple test=develop

* fix ut test=develop

* fix tests test=develop

* remove fp16 learning rate cast test=develop

test=develop

Stack OP has much higher CPU cost than expected in release mode.
Caused by DebugStringEx() in base class OperatorWithKernel. Actually
this issue occur for each OP which hasn't implement it's own
GetExpectedKernelType().

test=develop

* refine tensor
test=develop

* refine tensor
test=develop

* fix device_context log
test=develop

test=develop

test=develop

add MM_DNN inference test

Remove BinarySearch from Adam Op (CPU part)

test=develop

test=develop

test=develop

Enable fill_constant operator for a ngraph engine

Enable top_k operator for a ngraph engine