* First Commit.

* add some codes

* add elementwise loader

* fix code styles

* merge with develop

* add some changes both in elementwise and transpose

* add init operation in broadcast kernel.

* change codes according to pr suggestions about transpose file

* fix error for op-benchmark ci

* fix according to ci

* nullptr bugfix for XPU pg mode

Also a few kernels is added to xpu whitelist

* increase error msg length

* modify cmake file for cuda11.8 compile

* add op_library(fused_embedding_eltwise_layernorm_op DEPS bert_encoder_functor)

* Deleted mkldnn_inplace_pass code

* Fixed error with cmake

* Resolve conflicts

* add more unary op and square

* save fused_attention memory when dropout_rate = 0.0

* add ut

* fix ut bug

* fix fused_layernorm_residual_dropout_bias_test.cu

* generate static graph code of some ops by yaml

* fix the code-style of yaml

* fix the framework_ci for triangular_solve

* change the 'data_type' of scatter

* add the 'out: Out' of scatter_nd_add

* fix some transformer unitest

This reverts commit 592ed40b.

* fit paddle.matmul in fleetx.gpt

* add new tensor

* fix windows compile bugs

* fix ci bugs

* fix ci bugs

* fix ci bugs

* perfect according comment

* fix ci compile bugs

* add raw tensor

* fix ci bugs

* modify code by comment

* delete String

* fix codestyle

* add double complex<float> complex<double> dtype support for syevj_batched

* fix use_syevj flag for precision loss when input dtype of syevj_batch is complex128 in some case

* optimize eigh in different case

* fix missing ; bug

* fix use_syevj bug

* fix use_cusolver_syevj_batched flag

* fix: there are some bugs with trt 8.0

* fix:windows CI trt is too old

* Enable to print device info.

* Enhance the nan and inf checking for cpu.

* Implement a common print function.

* Unify the check of complex numbers.

* Rewrite the omp method.

* Count and print the number of nan and inf.

* Change the print content.

* Add unittest.

* move norm_utils.cu.h from fluid to phi

* remove norm_utils.h in fluid

* fix bugs and replace mutable_data with Alloc

* replace mutable_data with Alloc

* generate static graph code of some ops by yaml, test = develop

* fix 'take_along_axis' yaml style

* reset scatter/scatter_nd_add

* delete the comments of put_along_axis

* Add UT

* Support cross-step stream synchronization for standalone executor

* Fix typos

* Fix typos

* Update UTs

* chore: unify `SingleTensor`

* feat: dynamic check

* Add cast between int64 tensor and Paddle-TRT
* Add Unit testing.

* reshape

* typo

* remove header

* clean ir_pass_manager and fix map_depthwise_conv_to_conv_pass

* fix unitest timeout

* support custom_xpu

* update cmake to test xpu

* support custom_xpu, verify mechanism

* fix test_custom_relu_op_xpu_setup.py, test=kunlun

* fix FLAGS_init_allocated_mem

* cancel TIMEOUT property

* reset FLAGS_init_allocated_mem property

* add prelu

* mv fused_bias_dropout_residual_ln to fluid manual dir

* rm useless comments

* add cutlass act set in conv_elementwise_add_act_fuse_pass

* add suppport_tensor for code_gen to static graph

* support code-gen for int_array

* polish code

* fix bug of data_type

* move share_buffer kernel to phi

* fix ut

* add source file

* fix window links