* first commit

* fix code bugs in for_loop

* fix bugs in cuLoadAddStridedInputs.

* optimization for LayerNormBackwardComputeGradInput

* add unitest for validating the optimization

* fix windows ci error

* add fp16 and bf16 test

* update

* remove device_context include

* fix bug

* fix bug

* finish task

* add static_check and fix unittest.

* add int32/64

* Update test_cross_op.py

---------
Co-authored-by: NZhang Ting <Douyaer2020@qq.com>

* add all_gather and fix conflicts

* fix code format

* fix ut

* fix broadcast ut

* remove transpose infershape

* fix ci bugs

* fix ci bugs

* delete transpose infershape

* fix ci bugs

* fix ci bugs

* Replace paddle::experimental::DataType as phi::DataType

* restore custom_device.cc

* Add output defs for mode kernel

* fix bug

* add fp16/bf16

* add grad bf16

* test name

* platform::CUDAPinnedDeviceContext -> phi::GPUPinnedContext

* replace platform::TraceEventCollector

* remove fused_matmul from list

* add infermeta for fused matmul

* mkldnn->onednn

* fused softplus op + kernel

* remove extra attributes

* add missing handler

* change var name

* Add output defs for conv3d_coo distribute_fpn_proposals kernel

* fix

* add register of KthvalueKernel

add register of KthvalueKernel

* Update kthvalue_kernel.cc

* Update kthvalue_kernel.cu

* use python to generate cutlass code

* refine CommonConvKernelPart1, CommonConvKernelPart2

* remove useless code in generate_cutlass_code.sh

* add more config in conv2d_residual

* CommonCutlassConvKernelPart1 and CommonCutlassConvKernelPart2

* add group conv support in util.cu

* remove .sh

* refine name

* make name goodgit status!

* add fuse_alpha

* make code easy to understand

* mot fopen generate in py

* use python script to generate conv2d,group=1 cutlass code

* use const &

* use const & && use python script to generate conv2d/group=1 code

* Update interpreter_util.cc

* Update auc_kernel.cc

* Update auc_kernel.cu

* Update auc_kernel.cc

* Update auc_kernel.cu

* [XPU] add increment op.

* fix ci

* [XPU] add reduce_min_grad XPU kernel

* add unittest for reduce_min xpu op

* Update interpreter_util.cc

* Update bincount_kernel.cc

* Update bincount_kernel.cu

* Update interpreter_util.cc

* Update is_empty_kernel.cc

* Update is_empty_kernel.cc

* Add output defs for all_close all_raw  kernel

* Update interpreter_util.cc

* Add function node in phi_kernel for MKLDNN

* fix the bug in 'BuildInferVarKernelContext'

* add infer_varkernel_utils.cc

* fix the bug:the first two parametes of 'BuildInferVarKernelContext' can't be template variable

* change the code according to first review

* change the code according to first review

* change the mode of paddle_build.sh

* change 'infer_var_kernel_fn_' to 'get_kerneltype_forvar_fn_'

* add the error information

* fix NotFound infomation warning

* fix NotFound infomation warning

* fix NotFound infomation warning

* add register of yolo_loss_kernel.cc

* fix something

* Update paddle/phi/kernels/cpu/yolo_loss_kernel.cc

* delete something in interpreter_util.cc

* rollback

* add REGISTER of angle

* bug fix

* fix something

* Update paddle/phi/kernels/gpu/angle_kernel.cu

* Update paddle/phi/kernels/cpu/angle_kernel.cc

* Update paddle/phi/kernels/gpu/angle_kernel.cu

* Update paddle/phi/kernels/cpu/angle_kernel.cc

This reverts commit 3f4917f6.

* add numel INT64 register

* del numl

* add svd UNDEFINED register

* remove svd register

* remove operator.h from blas.h and remove paddle::framework::ExecutionContext

* remove the deps for GetBlas(exe_ctx)

* fix error

* [xpu-ernie3] support fp16 for full_like/stack/where xpu kernels

* [xpu-ernie3] support fp16 for full_like/stack/where xpu kernels