Test PaddingRNN on V100 GPU device.

Test configuration: large model, padding mode (which is the mode using recurrentOp), one GPU.
                   
GPU memory (MiB):   6414 (this PR)     vs   6837 (without this PR)
Speed (steps/s):         10.28 (this PR)    vs    9.89 (without this PR)
 

test=develop

* add check of executor, test=develop

* feature/auto_growth_allocator, test=develop

* add unittest of AlignedAllocator, test=develop

* try to turn on auto_growth to test on CI, test=develop

* fix segmentation fault in mixed_vector.h, test=develop

* add unittests, test=develop

* hash_op support int64 hash_size
* add corresponding UT

* remove ctr reader, all functions are satisfied in dataset

* make auc op compatible with 1 dim

test=develop

* fix int64_t

* update fill constant op unittest

* add empty line

2）add pip install coverage in Dockerfile.tmp
test=develop

* feature/buffer_shared_inplace, test=develop

* refine code, test=develop

* fix elementwise_add op cpu inplace and sum inplace bug, test=develop

* add unittest and debug log, test=develop

* fix parallel_executor scope bug, polish code, test=develop

* fix sum op, activation op, single_in_place_inference bug, test=develop

* remove kLocalExecScopeName, test=develop

* fix unittest,test=develop

* fix out_var first version bug, test=develop

* follow comments,test=develop

* test api example

* update python

* add sampcd_processor.py

* add if 0

* sort

* test paddle

* test paddle

* test paddle

* add whitelist

* change sampcd_processor.py

* change sampcd_processor.py

* change sampcd_processor.py

* add exit

* test=develop

* test=develop

remove dygraph.enable from __all__
hidden dygraph. profiler
add doc to dygraph. no_grad

* upgrade collective fleet api

* fix transform matrix bug, test=develop

* modify API.spec

* add random port

* Fix topk cannot handle 1D vector bug

Add path to handle 1D vector

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* refine code

test=develop
Signed-off-by: Nzhaoyuchen <zhaoyuchen01@baidu.com>

* test=develop, fix docker with paddle nccl problem

* test=develop, hide no_support api and add ut for it

* add_distributions_of_normal_and_uniform

* paddle/fluid/API.spec

* modify API.spec

* modified paddle/fluid/API.spec, test=develop

* modify paddle/fluid/API.spec, test=develop

* modify paddle/fluid/API.spec, test=develop

* fix some comment, test=develop

* modify API.spec, test=develop

* add comment for init function, modify hard code, test=develop

* modify API.spec, test=develop

* modify API.spec, test=develop

* make unit test function shorter, test=develop

* modify paddle/fluid/API.spec

remove pycpuinfo deps in core

* make it support float/int learning as input.

* enhance execution error info
test=develop

test=develop

* test=develop
support Tensor input for chunk_eval op

* test=develop
fix testcase for chunk_eval op

* test=develop
fix typos in nn.py

move the thread-check code from train_from_dataset to a single function
add UT for the thread check function

1. Since allreduce op has 4 reduce types, We split these four reduce types into four ops
2. We also refined the collective op code, e.g. we separated the collective op kernel into CPUKernel and CUDAKernel, and remove the device specified DeviceContext parameter in template as we already knew the target DeviceContext
3. We remove the newly added Collective op role to reduce the complexity of program and graph analysis

make fleet support mpi job submit directly.

* add not_been_used_vars to no_grad_set
test=develop