* copy kernel file to phi

* delete some code

* migrate uniform_random, test=kunlun

* fix input error, test=kunlun

* fix gpu register error, test=kunlun

* add include file, test=kunlun

* try fix error from CI, test=kunlun

* polish other PR

* fix CI-coverage error, test=kunlun

* config

* fix unittest

* zero init & cache & patch config

* add barrier to save and load

* add unittest

* refine cmake of framework

* add deps for dense tensor

* fix deps

* remove alloc(ctx)

* add depends on mkldnn

* add del dropout op pass to jit pe enigne

* add delete dropout test

* add runtime config in phi

* add runtime attr for op desc and op

* fix no proto error

* adjust opdesc set_attr impl

* try to remove conv_op extra attrs

* add init runtime attr map

* change extra header path

* fix runtime_attr

* fix trace_op

* fix bug of pass

* fix merge conflict

* fix dygraph attrs

* fix bug of pass

* fix dygraph bug

* fix unittest module

* delete extra attr default

* fix dropout kernel

* polish code

* fix extra output of instance_norm

* fix merge confilct

* fix op_desc bug

* add extra attr in yaml for conv3d_transpose

* don't remove extra input and output

* fix save_inference_model

* fix bug of batch_norm

* revert some change

* polish log

* polish code

* add code comment
Co-authored-by: NChen Weihang <chenweihang@baidu.com>

add constant folding pass， for some model，it will get less latency；

* add interpretercore

* refine backward program id

* add code

* refine program

* refine code

* create forward/backward_program by prog2graph2prog method

* test, do not care

* refine code

* refine code

* refine code

* test, do not care

* add interpretorcore

* add scope

* refine scope create method

* add jit for new_exe

* solve conflict

* delete unused code

* polish code

* polish code

* refine scope in inplace

* refine for datatransfer

* refine _rebuild_from_desc

* refine control eager deletion attr

* refine used_for_jit

* refine jit for infer

* op size0 use ori program

* polish code

* refine jit

* refine run_program_op ut

* refine inplace

* refine control

* refine graph helper

* refine control

* refine inplace

* refine buffer_share_inplace_pass

* polish code

* polish code

* refine usage for compilerProgram

* refine control

* test

* test core cache

* refine code

* refine io.py

* increase test_seq2seq timeout

* refine convert program

* refine interpretercore_cache release

* delete buildinplace

* refine partial_program && io

* refine code for io

* test

* test

* test

* [OpAttr]num_rows/num_colums of eye support Tensor type

* fix attr cast with long type

* remove fluid kernel and activate phi kernel

* fix parameter error

* transfer mkldnn part

* modify header file path

* fix compile error

* transfer special case

* fix lod setting and special case for layout setting

* add testcase and refine code

* update brpc version;test=develop

* [NPU] add run_program_op_npu

* add run_program_op_npu ut

* fix utest

* fix utest

* fix utest

* fix log

* fix random utest

* make tensor_util contains no cuda code

* refine isfinite

* revert ut

* move isfinite function to its op

* fix test

* fix compile

* std::isnan is not defined for int type on windows

* fix windows compile

* fix fp16

* fix rocm compile

* revert gradient node

* add quant pass

* Update scope.h

* typo

* Update dense_tensor.inl

* Add int8 support for matmul+elementwiae_add fuse

* Corrections after review and ernie test fix

* merge conv_concat_relu to conv_act

* fix typo

* extend unit test

* reuse existing gpd

* codestyle

* enforce mkldnn conv

* Fix random op depenency and lr_shedule bugs for standalone executor

* Fix CI errors

* Fix CI errors

* Fix CI errors

apply buffer_shared_inplace_pass and inplace_addto_op_pass pass to program in Standalone Executor (#45085)

* apply inplace addto in python apply_pass

* fix

* apply inplace pass for program

* skip feed and fetch var

* fix block_desc.move_from

* fix block desc

* alltoall remove inplace

* fix

* change to async mode for xpu multi-card training in static graph mode

* minor bugfix

* irrelevant. move to another pr

* move change to other pr

* fix stream issue

* fix 'stream not meet with current context' error

* fix branch diverge, test=kunlun

* fix infer tans scop

* fix infer trans scope

* fic infer trans scope

* fic infer trans scope
Co-authored-by: Ndingjiawei <327396238@qq.com>

* [OpAttr]Add SupportTensor for OpMaker

* fix typo

* fix code style

* add SupportTensor for concat op

* add unittest for register Tensor

* add shape checker and split attribute

* move check finite and unscale kernel into phi

* move infershape into phi

* move update_loss_scaling kernel into phi

* remove original kernels

* move update loss scaling infershape into phi

* add header for xpu and npu

* solve coverage failed

* fix npu test failed

* remove mutable data in cu file

* fix new executor failed

* add valid check for meta tensor output

* convert multihead to oss

* fix:bug

* fix:delete const cast

* fix:don't support bias_qk

* add vit pass

* fix:convert bug and add preln_residual_bias

* support length=-1

* add UT for convert

* add no_bias_qk support for gpu_multihead_op

* delete infer_shape depends on bias_qk

* oss just can be used in T4 and A*

* fix:change api for ROCM CI

* [Auto Parallel] Move the distributed info from python to c++

* [Auto Parallel] Add dist_attrs for VarDesc and OpDesc

* [Auto Parallel] Add the lost file

* [Auto Parallel] Make the dist attr be unique_ptr

* [Auto Parallel] Add the proto conversion

* [Auto Parallel] Improve the proto support

* [Auto Parallel] Fix the bugs for adding a device or a link

* [Auto Parallel] Add the C++ ProcessMesh and DistributedMapper

* [Auto Parallel] Improve the impl of these dist attrs

* [Auto Parallel] Pybind11 ProcessMesh and DeviceMesh

* [Auto Parallel] Fix the unittest problem

* [Auto Parallel] Explicitly add the src file for auto_parallel target

* [Auto Parallel] Add the proto depedency explicitly

* [Auto Parallel] Fix the cmake bug on windows and mac

* [Auto Parallel] Remove the pybind11 header file in process_mesh.h

* [Auto Parallel] Remove unused codes

* [Auto Parallel] Check whether the dist attr is null

* [Auto Parallel] Implement the assign operator for OpDesc explicitly

This reverts commit 84bf5c31.

* add cached_serialize_str_

* support program hash

* add sha

* add ut

* use hash_str only for new_exe

* fix attr order