* add scale gather sum

* refine CUDA_ATOMIC_WRAPPER ADD for bf16

* add gather unittest

* solve conflict

* add scale uinttest

* add sum unittest

* solve conflict

* refine gather unittest

* refine unittest

* remove SetAllocationForOutputTenosr

* add place param for copy kernel

* recover SetAllocationForOutputTenosr

* polish code

* fix empty_dev api bug

* remove reseting dtype and layout for output in executor

* fix merge bug

* [Phi] Add ClearHolder when re-alloc on new place in DeviceContext

* fix hostAlloc

* remove setting output allocation

* remove full_kernel_impl.h

* fix bug of xpu full_like
Co-authored-by: NAurelius84 <zhangliujie@baidu.com>

* move uniform_random to phi

* fit selected_rows

* replace mutable_data

* support kps backend and compile

* resolve conflict

* fix kps backend trans

* test in xpu2 device

* remove dummy kernel

* optimize mergeadd for sparse_adam,*test=kunlun

* optimize mergeadd for sparse_adam,*test=kunlun

* optimize mergeadd for sparse_adam, *test=kunlun

* add multi input for infer_shape

* support multi output for infershape

* fix split bug

* fix bug of concat

* support vector<MetaTensor*> in infrt

* fix bug

* [Phi] Migrate logical_and/or/not/xor into Phi

* fix unittest

* fix function name

* add reducer

* optimize group norm forward

* use vectorized optimization

* add scalar calculation code

* optimize code

* vectorize lamb kernel

* remove flags, add ut

* remove useless codes

* refine code, add param order

* graph engine demo

* upload unsaved changes

* fix dependency error

* fix shard_num problem

* py client

* remove lock and graph-type

* add load direct graph

* add load direct graph

* add load direct graph

* batch random_sample

* batch_sample_k

* fix num_nodes size

* batch brpc

* batch brpc

* add test

* add test

* add load_nodes; change add_node function

* change sample return type to pair

* resolve conflict

* resolved conflict

* resolved conflict

* separate server and client

* merge pair type

* fix

* resolved conflict

* fixed segment fault; high-level VLOG for load edges and load nodes

* random_sample return 0

* rm useless loop

* test:load edge

* fix ret -1

* test: rm sample

* rm sample

* random_sample return future

* random_sample return int

* test fake node

* fixed here

* memory leak

* remove test code

* fix return problem

* add common_graph_table

* random sample node &test & change data-structure from linkedList to vector

* add common_graph_table

* sample with srand

* add node_types

* optimize nodes sample

* recover test

* random sample

* destruct weighted sampler

* GraphEdgeBlob

* WeightedGraphEdgeBlob to GraphEdgeBlob

* WeightedGraphEdgeBlob to GraphEdgeBlob

* pybind sample nodes api

* pull nodes with step

* fixed pull_graph_list bug; add test for pull_graph_list by step

* add graph table;name

* add graph table;name

* add pybind

* add pybind

* add FeatureNode

* add FeatureNode

* add FeatureNode Serialize

* add FeatureNode Serialize

* get_feat_node

* avoid local rpc

* fix get_node_feat

* fix get_node_feat

* remove log

* get_node_feat return  py:bytes

* merge develop with graph_engine

* fix threadpool.h head

* fix

* fix typo

* resolve conflict

* fix conflict

* recover lost content

* fix pybind of FeatureNode

* recover cmake

* recover tools

* resolve conflict

* resolve linking problem

* code style

* change test_server port

* fix code problems

* remove shard_num config

* remove redundent threads

* optimize start server

* remove logs

* fix code problems by reviewers' suggestions

* move graph files into a folder

* code style change

* remove graph operations from base table

* optimize get_feat function of graph engine

* fix long long count problem

* remove redandunt graph files

* remove unused shell

* recover dropout_op_pass.h

* fix potential stack overflow when request number is too large & node add & node clear & node remove

* when sample k is larger than neigbor num, return directly

* using random seed generator of paddle to speed up

* fix bug of random sample k

* fix code style

* fix code style

* add remove graph to fleet_py.cc

* fix blocking_queue problem

* fix style

* fix

* recover capacity check

* add remove graph node; add set_feature

* add remove graph node; add set_feature

* add remove graph node; add set_feature

* add remove graph node; add set_feature

* fix distributed op combining problems

* optimize

* remove logs

* fix MultiSlotDataGenerator error

* cache for graph engine

* fix type compare error

* more test&fix thread terminating problem

* remove header

* change time interval of shrink

* use cache when sample nodes

* remove unused function

* change unique_ptr to shared_ptr

* simplify cache template

* cache api on client

* fix

* reduce sample threads when cache is not used

* reduce cache memory

* cache optimization

* remove test function

* remove extra fetch function

* graph-engine data transfer optimization

* support graph_split load&query

* remove logs

* change shards to pointer vector

* use inference

* remove test code

* renorm op

* simplify renorm op

* recover local changes

* recover renorm op kernel

* fix init

* add blanklines in renorm doc

* fix import

* fix import

* add renorm to init.py

* merge

* move index_sample op

* Delete api.h

* Delete api.cc

* fix

* remove logs

* recover infer shape of grad

* recover changes

* change shape

* fix label

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix
Co-authored-by: NHuang Zhengjie <270018958@qq.com>
Co-authored-by: NWeiyue Su <weiyue.su@gmail.com>
Co-authored-by: Nsuweiyue <suweiyue@baidu.com>
Co-authored-by: Nluobin06 <luobin06@baidu.com>
Co-authored-by: Nliweibin02 <liweibin02@baidu.com>
Co-authored-by: Ntangwei12 <tangwei12@baidu.com>

* move size, erfinv, pixel_shuffle infershape to phi

* fix erfinv infermeta

* init grid_sampler with mode=bilinear

* solve error

* rm fill constant

* rm head

* change block size

* change block size

* optimize

* apply existing config

* add align for WorkQueue

* add spinlock

* merge develop

* merge

* Add EventsWaiter

* Revert "Add EventsWaiter"

This reverts commit e206173aa9be7401b83a53581627bfaf557c8fb2.

* Add host_trace_level env variable

* Revert "Optimize perf of softmax_with_cross_entropy (#39553)"

This reverts commit bbe5228c.
Co-authored-by: Nliutiexing <liutiexing@google.com>
Co-authored-by: NZzSean <18818272991@163.com>

* add align for WorkQueue

* add spinlock

* merge develop

* merge

* Add EventsWaiter

* Revert "Add EventsWaiter"

This reverts commit e206173aa9be7401b83a53581627bfaf557c8fb2.

* add log for Executor

* Profile Allocators

* Profile Allocators

* adjust interface

* remove lock for set

* fix
Co-authored-by: Nliutiexing <liutiexing@google.com>

* fix incorrect dims settings

* sparse conv3d

* fix out dims

* test performance

* test large shape success

* opt scatter, double performance

* test float16

* remove profiling code

* remove pten

* opt code lines

* correct boundary judgment

* only cpu

* test ci

* test ci

* remove the including paddle/fluid header; extract the conmmon function

* opt code lines

* use DenseTensor::data() instead of mutable_data

* return rulebook for backward

* specify layout

* rename:conv -> sparse_conv3d

* [Phi] move truncated_gaussian_random, copy kernels

* [Phi] move truncated_gaussian_random, kernel register

* [Phi] move truncated_gaussian_random, delete useless codes

* add some trt layers

* trtOpConverter pass ok

* add comments

* add constraints to some attrs in the pd_lower_to_trt patterns

* update constraint

* fix code style

* update pass name

* update code style

* change .hpp.inc to .cc.inc in mlir_add_rewriter

* update;test=cpu-py3

* rename pten_utils to phi_utils

* rename pten_utils target

* rename Pten to Phi

* replace pten with phi

* resolve conflict

* add align for WorkQueue

* add spinlock

* merge develop

* merge

* Add EventsWaiter

* Revert "Add EventsWaiter"

This reverts commit e206173aa9be7401b83a53581627bfaf557c8fb2.

* update HostTracer

* fix

* update

* update
Co-authored-by: Nliutiexing <liutiexing@google.com>

* refine bf16 amp-o1 logic

* refine amp GLOG

* refine unittest

* refine unittest

* remove empty kernel in fluid and adjust the param of empty dev_api

* polish code

* revert fluid empty kernel

* fix selected_rows bug in C++ API

* add optional for C++ APIO

* data transform support optional

* remove data transform for optional vector<Tensor>

* adjust some format of funtcion

* fix empyt bug

* add new profiler components

* fix bug

* [KP] Unify .cu and .xpu files with .kps files

* fix CI bug in GPU and modify the list

* fix conflict

* modify the date

* [Phi] Add ClearHolder when re-alloc on new place in DeviceContext

* fix hostAlloc

* foix inferRT unittest

* remove dev_ctx ptr

* fix pylayer problem with amp

* add ut

* refine code