Add key_padding_mask and attn_mask in sparse_attention Api

1.Key padding mask is a tensor with dimensions [batch_size, seq_len], and attention mask is a tensor with dimensions [seq_len, seq_len]. The data types of the two masks are consistent with Q, K, and V, which are float32 or float64. If the value in Mask is 0, it means that the position needs to be masked.

2.The changed files are mainly paddle/fluid/operators/sparse_attention_op.cu and python/paddle/fluid/tests/unittests/test_sparse_attention_op.py. sparse_attention has three parts: sddmm, softmax, and dsd. Adding the mask operation only needs to modify the softmax. It has no effect on the other two parts. In addition, in order to test the mask function, related tests has been added.

* Add the transformop parameter in TensorReduceFunctorImpl

* Reduce reshape kernel functions in pten

* delete notes

* fix bugs when compile

* modify register name

* fix compile bugs

* Add float16 type for scatter op.

* Add fp16 test for scatter op.

* Add int and int64 support for scatter_grad on gpu.

* Add int and int64 for check_variable_and_dtype routine.

* Minors.

* Code format.

test=document_fix

As the title.

* replace with pten kernel in cast cuda compute and remove unused codes

* rm unused header file

* replace CastCUDAOpKernel with CastOpKernel

* add map_matmul passes to quant2_int8_mkldnn_pass

* fix fc+act fuse (activation scale)

* ci fix, c++17 structured bindings not available

* fix ci static check

* add conv_gelu_mkldnn_fuse_pass

* add post ops

* Reduce reshape kernel functions in pten

* delete notes

* fix bugs when compile

* reshape+transpose+matmul_v2

* in_name->input_name

* fix pr-ci-static-check

* add Logit API

* add unittest

* conflict

* pull conflit

* pull conflit logit

* fix unititest

* fix code style

* update docs style of

* update en doc

* fix docs en style

* fix docs en style1

* fix docs en style2

* fix docs en style3

* fix docs en style4

* fix docs en style5

* fix docs en style6

* fix docs en style7

* fix docs en style8

* update by review

* fix nan bug

complement deps of cmake files under WITH_CINN compilation

* fix

* modify log

* fix batch_size

* add as_complex and as_real op

cinn_launch_op： cache scope and place on CinnLaunchContext to skip duplicate alloc/free callback construction

* refine the kernel register impl

* fix cmake and symbol error

* remove overload marco

* polish details

1. add a subdirectory named `cinn` in `paddle/fluid/operators` directory and move releated files into it
2. seperate CinnLaunchContext class from `cinn_launch_op.h` and put it in a  new independent file named `cinn_launch_context.h`, so that it can be included by others clearly.

* add alias kernel name

* modify code as suggestions

* save temp

* add unittest, test=develop

* fix ci error, test=develop

* fix grad accuracy error, test=develop

* fix unused error, test=develop

* fix compilation error on Windows, test=develop

* add unittest, test=develop

* modify by review comment and add lerp_

* fix inplace api, test=develop

* fix inplace api, test=develop

* fix coverage error, test=develop

* add boardcast_sub

* add boardcast_sub

* fix CUDA Graph H2D bug again

* fix no return bug

* quantize slice op

* correct test

* fix code formatting

1. Modify git tag for CINN
2. Support compile option "-DWITH_CINN=ON, -DWITH_TESTING=OFF"

* fix reshape move storage error

* remove needless set type

* alloc tensor by shared storage

* refine structure for cuda and rocm

* update

* update

* update

* update

* use sync h2d copy

* use stream safe allocator in memcpy_h2d

* remove wait

* add guard