* Add XPU plugin to support the customized ops or improve the performance of the fusion ops based on hand-written xpu micro kernels.

* refine README.md

* suport optional input in new_ir

* polish code

* add coverate test

* update

* update

* add unitest

* remove reduplicate code

* set test timeout

[XPU] Fix the topk, set_value ops that using temporary tensors avoiding the memory overlaps during multi-stream inference (#54851)

* support auto-gen concat

* fix bug in legacy_backward.yaml

* fix bug in get_expeceted_kernel_type

* configure elementwise_mod op_version

* support auto generate for static op elementwise_mod

* [CustomDevice] release device manager in py::atexit

* fix hip_version macro

* update

* update

* configure elementwise_floordiv op_version

* support auto generate for static op elementwise_floordiv

* update unity_build_rule.cmake

* fix the static op generation for group_norm

* fix bug

* fix bug

* Update op_compat.yaml

* add lerp bf16 support

* fix bug

* Update test_lerp_op.py

modify the input dtype

* modify the test_lerp_op.py

* Update test_lerp_op.py

* fix bug of import

* add user_defined_grads

* Update test_lerp_op.py

* fix bug of grad

* fix bug of grad

* fix bug of grad

* add the check for bfloat16 dtype

* add linear_compress API

* fix_fetch_op_and_null_type_bug

* fix compile bug

* add test case

* refine program translator

* fix warning: not override

* fix bug

* merge new modifications

* modify by reviews

* resolve conflicts

* resolve conflicts

* fix

* fix

* fix conflicts

* add unittest for special op transcriber

* set cpu as default backend

* modify by reviews

[prim] modify multiply_double_grad composite rules to solve paddlescience poiseuille_flow.py problem (#54908)

* modify eular_beam

* modify matmul infermeta

* add test

* modify timeout

* modify mutiply_double nullptr

* modify tanh_triple_gradnode create

* static graph autogen code for expand

* fix bug

* fix bug

* fix bug

* fix bug

* fix bug

* fix bug

* update

* refactor build context

* fix bug

* polish code

* change func name

* Fix compiling on XPU related to MPTypeTrait.

* Unify the use of MPTypeTrait.

* Fix compiling error.

* add kernel dialect

* change DenseTensorTypeStorage to DenseTensorType

* add test case`

* add first pd_op to kernel dialect

* lower pd op to kernel dialect

* update

* update

* remove useless code

* add attrite print test

* fix bug

* update

* update

* update

* update

* polish code

* fix bug

* polish  code  and add python test

* add test

* fix test error

* add env flag

* fix bug

* revert test env

* change cc_test_old to cc_test

* fix build_static bug

* fix type test error

* udpate cmake

* disable test in windows

* update

* update

* fix bug

* split file

* fix conflict

* polish code and fix conflict

* support place transformer

* finish bug

* add gpu flags

* fix with cuda macro

* add fetch kernel

* support fetch var in new ir

* fix bug

* polish code

* change array equal to np.testing

* support feed in new ir

* update

* fix bug

* try to hack combine op

* add scope guard

* revert atan2 op

* add scope guard

* update

* polish code

* update

* refactor build kernel context

* fix unitest bug

* polish code

* use original order

* remove useless code

* polish code

* fix bug

* style

* more

* update ctest

* Update legacy_backward.yaml

* Update legacy_ops.yaml

* Update legacy_ops.yaml

* update

* update

* update for move

* add clone for inpalce

* fix name

* add inplace pow

* fix typro

* add note

* fix typro

* fix typro

* fix bug

* fix test error

* add type error test

* adjust indentation

* add int_with_ll to conv

* fix bugs when output_size is specified for conv2d_transpose

* add pos_weight in kernel

* fix unittest

* fix xpu

* fix bce unittest, change infermeta order

* [ROCM] fix cupti, hipcub

* update

* update

* delete swish_raw

* fix

* Update activation_kernel.cc

* fix

* base rule

* add sharidng merge

* add sharidng axis merge

* define unified data class for inferencing dist_attr

* test wrap DistTensorSpec in dygraph mode

* matmul main logic done

* define unified data class for inferencing dist_attr

---------
Co-authored-by: NYichen Zhang <zhangyichen03@baidu.com>

* add all_to_all phi operator, kernel, api

* add all_to_all ut

* tinyfix

* add kernel dialect

* change DenseTensorTypeStorage to DenseTensorType

* add test case`

* add first pd_op to kernel dialect

* lower pd op to kernel dialect

* update

* update

* remove useless code

* add attrite print test

* fix bug

* update

* update

* update

* update

* polish code

* fix bug

* polish  code  and add python test

* add test

* fix test error

* add env flag

* fix bug

* revert test env

* change cc_test_old to cc_test

* fix build_static bug

* fix type test error

* udpate cmake

* disable test in windows

* update

* update

* fix bug

* split file

* fix conflict

* polish code and fix conflict

* support place transformer

* finish bug

* add gpu flags

* fix with cuda macro

* update

* add scope guard

* polish code

* add fetch kernel

* support fetch var in new ir

* fix bug

* polish code

* change array equal to np.testing

* support feed in new ir

* fix bug

* try to hack combine op

* add scope guard

* revert atan2 op

* polish code