* extract sub-graph

* graph-engine merging

* fix

* fix

* fix heter-ps config

* test performance

* test performance

* test performance

* test

* test

* update bfs

* change cmake

* test

* test gpu speed

* gpu_graph_engine optimization

* add ssd layer to graph_engine

* fix allocation

* fix syntax error

* fix syntax error

* fix pscore class

* fix

* recover test

* recover test

* fix spelling

* recover

* fix

* split reduce_kernel

* rm reduce_kernel in cmake

* split reduce_grad kernels

* fix cmake build error

* format code

* fix standalone_executor_test error

* add gpupsutil and afsclient; test=develop

* refactor heter comm kernel

* update. test=develop

* update calc_shard_offset. test=develop

* update xpu kernel. test=develop

* update args of calc_shard_offset

* update. test=develop

* remove customGradMerger

* update. test=develop

* update. test=develop

* fix. test=develop

* update. test=develop

* update. test=develop

* update optimizer kernel

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* update. test=develop

* fix. test=develop

* fix. test=develop

* add optimizer kernel. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix kunlun not support size_t. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* update hashtable. test=develop

* update. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* update. test=develop

* update. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* template init. test=develop

* hashtable template init. test=develop

* fix. test=develop

* fix. test=devlop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix hashtable_kernel. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop

* fix. test=develop
Co-authored-by: NWorgenZhang <frank08081993@gmail.com>

* add mixed-precission support for ipu

* restore cast_model_to_fp16 api

* update UTs

* regist elementwise_xxx

* executor perf statistics

* fix ut

* fix ut

* fix ut

* add ut

* add ut

* Add UT

- Added missed data_layout

- Added missing conversions

- NDHWC added

- NDHWC support in data_transform

- another fix

- condddate change

- fix

u- fix

- fix

- fix

- fix

- fix

- fix to hack

- compilation fix

- fix to automatic merge

* - reduced UT

* - fix

* - lint

* - fix to lint

* Change tensor name to match activation

* declare fc_eltwise_add pass

* merge conv_eltwise refactor PR

* first compilable draft

* unittest feedback tools

* Fuse pass tester

* Move IsReachable() to shared file

* 100% coverage of fuse_pass_tester.cc

* register pass

* Add bias node

* Improve unit tests / remove bias node from pattern

* improve fc_eltwiseadd_unittest

* cancel eltwise_add fuse if act is already fused

* Add elementwise_input scale

* Residual MVP

* Add new FC attrs

* Add more test cases

* Add missing op attrs

* Adapt code to new Elementwise pattern

* reuse existing fcpattern

* improve code style

* remove unused arguments

* fix typo

* remove whitespace

* remove int8 related code

* Remove attributes from base ops

* style

* style check

* Remove input from base op

* Set attribute during fuse

* ut timeout

* download and test model

* DRY

* apply feedback from review

* Style check

* fix typo

* cosmetic changes

* explicitly set residual as output

* VIT-OCR accuracy check

* trigger CI

* remove whitespaces

* fix missing data file

* add mkldnn int8 pass [step3]

* Add test for compute_propagate_scales_mkldnn_pass

* update pass

* update api comment and python api
Co-authored-by: Nwozna <joanna.wozna@intel.com>

* added shuffle_channel bf16/fp32 fwd kernel

* added missing files

* CI fix

* changed from pten to phi

* tmp save

* added reviewers suggestions

* fix for test

* the one ps proto

* the one ps proto

* fix

* fix

* fix

* fix windows ci

* fix windows ci

* add dependency

* add dependency

[XPUPS]add support for kunlun2
Co-authored-by: NWorgenZhang <frank08081993@gmail.com>

* remove stack_grad infershape

* fix bug of output with null

* fix bug

* optimize hbm

* format

* format

* remove old custom op placetype

* replace dist  placetype using

* add with gpu macro

* fix mutable_data error

* fix set value error

* add comment

* perform SlotRecordInMemoryDataFeed feedvec;test=develop

* add new method for custom double grad

* add tanh double grad unittest

* change year

* revert tensor init method

* [KP]fix bug when TruncatedNormal cannot fall back in cpu

* delete useless comment

* delete useless comment

* add mkldnn int8 pass

* add mkldnn int8 pass

* update pass

* update name

* update name

* fix test

* fix fleet bind

* update name

* update name

* fix test

* fix gpups wrapper

* remove Push/Pull/Load/Save with context in client and wrapper base class

* fix

* fix

* remove some interface

* fix

* remove

* code style

* recover

* fix

* remove code unused

* remove some unused table & accessor & CommonDenseTable => MemoryDenseTable

* fix

* fix

* fix

* recover

* remove unused code

* recover unittest

* fix

* remove

* fix

* remove code unuseful

* remove

* fix

* recover

* remove
Co-authored-by: Nesythan <esythan@126.com>

* Using the maximum workspace_size of all alogirhms to limit the workspace size in exhaustive search mode.

* Use the system cudaMalloc and cudaFree to allocate workspace during searching.

* Enable switch of two kind of workspace setting methods.
Co-authored-by: NLiu Yiqun <liuyiqun01@baidu.com>

* fix bug that no thread is waked up when adding task to threadpool

* fix typo

* afs wrapper

* format

* format

* macro

* Profile Executors

* update

* fix ut

* fix names

* update

* update

* momentum support l2decay for xpu. test=kunlun

* fix include file. test=kunlun

* fix cmake for device_worker. test=kunlun

* remove paddle_ipu shared library

* fix unique_name

* fix bug of data transform in inference executor

* fix bug

* enable new executor by default

* enable stream safe allocator

* test=document_fix;test=coverage

* do not use scope in op kernel

* fit empty program for new executor

* fix communication depend

* fix test_sync_batch_norm

* skip unsupported place

* refine datatransfer

* fit for dirtributed program

* fix dependencpy

* fix some ut

* DRY

* change nodes names

* add const prefix

* change asX to as_x in all files

* add dropout slice yaml

* remove useless code

* fix infer shape error

* skip infrt compile for dropout

* add concat_grad kernel

* fix error

* remove comment code

* fix outs nullptr error

* change to phi header

* add concat_grad declare for standalone_executor_test

* paddle inference support new quant_model

* [KP] fix bug in phi static graph mode

* modify the useless code

* update name

* update name

* fix test

* fix fleet bind

* update name

* update name

* fix test

* fix gpups wrapper

* remove Push/Pull/Load/Save with context in client and wrapper base class

* fix

* fix
Co-authored-by: Nesythan <esythan@126.com>

* fix mac c++ version

* update

* fix apple systems