* blockReduce opt

* launch threads align to warpSize

* reduce unnecessary shared memory for broadcast reduced value

* vectorize SoftmaxKernelWithEltadd

* add fp16 constrain

* test=develop

As the title

* Add a StatValue class in the backend to represent a stat.
* Add a singleton StatRegistry to maintain the collection of stats.
* For the sake of code neatness, we only support type of int and float, which can cover most of the scenarios.

Some big models can timeout on Windows CPU machine. I added some timeout properties.

This commit fixs the compiling bug regarding unique_ptr of IOptimizationProfile.

IOptimizationProfile has protected dtor and is controlled by TensorRT
internally. Application shouldn't delete the pointer of IOptimizationProfile.
See TensorRT document: https://docs.nvidia.com/deeplearning/sdk/tensorrt-api/c_api/classnvinfer1_1_1_i_builder.html#a9ac47e100454151d8206ac91d543299a
test=develop

* parallel move shared data test=develop

* test=develop

* update readme of 1.8.2;test=document_fix

* test=develop;test=document_fix

* test=develop;test=document_fix

Add TSM as ProgramTranslator Unit Test. The TSM code is referred from PaddlePaddle/models#4229

* test=develop

* test=develop

* fix bug

* test=develop

* test=develop

* add MobileNet unittest test=develop

* fix cudnn random test=develop

* fix sync barrier with barrier monitor, test=develop

* add summary_env, test=develop

* update issue template, test=develop

* refine link, test=develop
Co-authored-by: Nroot <root@yq01-gpu-255-129-15-00.epc.baidu.com>

* enable load_program_state run in imperative mode; test=develop

* remove useless code; test=develop

[Dy2Static]Convert var.shape stmt and Convert the return variables of Tensor-dependent 'if' staments to Tensor if it not  (#24911)

* Support int and long: int or long -> six.integer_types. 

* Modify test_tensor_shape: fix bug and modify comment. 

* Support convert_var_shape to convert var.shape stmt

* Modify code in ifelse_simple_func.py because don't support return non-Tensor in Tensor-dependent 'if' stament currently. 

* Convert the return variables of Tensor-dependent 'if' staments to Tensor if it not. test=develop

* use allow list instead of white list, test=develop

* reduce include, test=develop

Support CMatchAucCalculator based on CMatchRankAucCalculator with a new parameter ignore_rank

[Dy2stat] decrease the batch size to decrease GPU usage.

* windows publish package scripts,test=develop

* windows publish package scripts,test=develop

* windows publish package scripts,test=develop

fix bug in CUDA_NVCC_FALS and CMAKE_CUDA_FLAGS

* test=develop, add log message in the function UpdateDllFlag

* test=develop, add the test

* remove gmock from ut

test=develop

* coverage enabled for r+t+m fuse pass

test=develop

* Move function 'convert_len' to file convert_operators.py 

* Support that for statements are transformed to while statements.

* Fix bug: raise None -> return None. 

* Support variable loaded and created in loop.

* Use int64 in Py2 and Py3 in function to_static_variable.

Fix the bug for elementwise_div op, when the first var is scalar; Use the shape 1 replace the -1 in shape.

[Dy2stat] Add word2vec as unittest

Add the support the 5d,6d tensor support for the reduce ops;
Add the same time, the compile time, it was 22 minutes, it was 21 minutes after fixed.