* [dy2static-bugfix] fix backward gradient aggregation bugs
1. Yolov3 and Yolov5 all face the same problem.

* remove set_device

* code review fix

* add logspace yaml

* update by comments

* resolve test framework conflicct

* support xpu multi-card infer

* add ut

* clean code

* clean code

* fix

* fix

* fix

* fix

* polish namespace

* change static_tensor_operants

* polish namespace

* support add, subtract, divide

* add unit test

* polish unittest

* fix cmake error

* polish unittest

* [XPU] update xccl to 1.0.8

* update xdnn. add uint8 for concat and split.

* update xdnn to 20230215.

* [XPU] add group_norm kernel

* update

* add xpu sin, cos, randint, linspace kernels

* update

* update

* move layer_norm_kernel.cu.h to phi

* fix bugs

* fix namespace

* fix bugs

* fix CI-Windwos

* replace mutable_data

* fix bugs

* fix bugs

* move variable_utils from phi_api_utils to fluid

* fix coment

* update include

* fix bugs

* fix bugs

* fix bugs

* fix bugs

* fix bugs

* update

* update

* fix CI-Windows-OpenBLAS

* fix bugs

* fix bugs

* fix bugs

* update include

* move variable_utils to phi_utils

* fix namespace

* map output from composite rule to origin op

add mean layer_norm dropout op map

add input map check

composite softmax support input shape []

* composite softmax support shape []

* polish log

* solve conflict

* polish code

* polish op map output

* add check dtype

* move profiler

* add file

* fix mac compile bugs

* fix ci bugs

* fix mac bugs

* fix ci bugs

* fix compile bugs

* perfect code according comment

* [CUSTOM]custom device add black_list

* change log level

* fix some issues

fix X

remove TensorCopy

codestyle

add fluid memory header

fix symbol

fix cmake

fix cmake

fix context

fix header

fix place

fix context

fix context

fix context

fix code

fix custom context

fix custom context

fix copy

fix data_transform

fix style

remove changes of custom

fix scalar

* expand mix_precision to custom_device

* fix bug

* fix bug

* fix comment

* fix DEFINE bug

* polish namespace

* change static_tensor_operants

* polish namespace

* first commit.

* a little changes

* add some changes for get vec_size efficiently

* fix bugs

---------
Co-authored-by: Nzhangbopd <1299246947@qq.com>

* delete axis of fmin

* fix bug

* Tensor support void* data() function

* add unittest

* add selectedRows unittest

* polish unittest

* polish unittest

* polish unittest

* polish unittest

* add xpu adagrad and where_grad kernels, test=kunlun

* add xpu pool3d kernels, test=kunlun

* support size 0 dot input

* prevent div 0 in grad

* add unittest

* remove unnecessary vlog

* add unittests

* init commit

* fix tensor operator*

* fix compile bug

* bug reproduce

* update commit

* polish codes

* fix compile bug

* test begin

* test begin

* compile finish

* restore origin composite_backward_api

* pass local CI

* fix merge error

* fix merge error

* change py_test from GPU->CPU, test custom op

* polish codes, modify prim unittest

* modify prim unittest

* determine phi_tensor_operants location

* polish codes

* add header file

* solve windows unresolved symbol

* fix some CI error

* add overload defination

* fix CI inference and Windows

* polish codes according to reviewers' opinion

* polish codes according to reviewers' opinion

d_bias are nullptr. Modify the code style of full_kernel.cc. Add new data
type for concat, elementwise_add, gather, scale, scatter ops. test=kunlun

* add dim check in scatter

* add check in scatter.cu

* add unittest

* remove unnecessary log and comment

---------

Co-authored-by: RedContritio <>

* remove AllocatorFacade in phi

* fix include

* fix bugs

* rm gradient_accumulator in phi

* update

* decouple strided_memcpy

* move strided_memcpy

* move strided_memcpy to phi

* fix namespace

* update

* fix gpu compile bugs