* Quantize nearest_interp and nearest_interp_v2

* Check if avx_core supported

* Add depthwise_conv2d to supported quantization list

* change 'math' to 'math_kernel'

* fix compile bugs

* merge develop

* fix compile bugs

* move cpu_impl of elementwise kernel to new directory

[Unify Tensors PR #3]Port framework::Tensor members & interfaces to pten::DenseTensor, test=allcases (#38473)

* Added shared_ptr<Allocation> member & corresponding interfaces to Storage

* Removed original pten::Allocation from Storage and adjusted the interfaces accordingly

* Fixed issues with storage offset

* Used place to malloc allocation for TensorStorage

* [Unify Tensors PR #3]Ported framework::Tensor interfaces to pten::DenseTensor

* Fixed issues with place

* Added comments

* Moved mutable_data with stream argument to DenseTensor

* Added set_offset interface

* Fixed CI issues,test=allcases

* [Unify Tensors PR #4] Port LoDTensor interfaces to DenseTensor

* Reverted changes too pten_layout() interface

* Removed friend classes

heter context support dynamic mf dim

* add mul_gru_fuse_pass ut

* update ut

* update ut

* update ut timeout setting

* update ut

* [MLU]support calling mlu op from python interface

* [MLU]fix

* fix

* [mlu]fix mlu_places

* [mlu]fix required mlu

* fix

* [MLU]fix tensor copy

* [mlu] fix MLUPlace call path

* unify data layout

* fix test_transfer_layout error

This PR adds runtime flags run_kp_kernel, which choose which op to run for xpu2. There are two: dynamic linked and built from kp.

* [Auto parallel] Make the id of var and op unique

* [Auto Parallel] Rename back dist_context to distop_context

* dynamic shape clone supported

* fix wait for tiexing

* fix work2vec model. new_exe support EOF Exception in ReadOp now

* remove offset in storage

* revert api change

* fix custom op slice bug

* fix mutable_data error

add hashtable dynamic mf support

add dynamic mf size api

* auto parallel sharding base

* chmod

* add unitest

* set unitest cmake dist label

* revise code according to rewiew

* chmod

* fix reshape move storage error

* remove needless set type

* alloc tensor by shared storage

* Utilize StreamSafeCUDAAllocator to support fast GC in new executor

* Fix compile error for Windows and ROCm

* Fix compile error for Windows

* Modify UT stream_safe_cuda_alloc_test

* Modify UT stream_safe_cuda_alloc_test

* Rewrite fast GC

* Rewrite fast GC

* Fix compile error for BOOST_GET_CONST

* Fix compile error for BOOST_GET_CONST

* Changes default stream for StreamSafeCUDAAllocator

* Fix a small CI error

* Remove some redundant code

* Fix conflict

* Fix compile error for ROCm

* Fix Windoes CI error

* Fix CI error

* Remove some unnecessary code

* Fix CI error

* Add UT for fast GC

* Fix CI error

* add device-agnostic stream class

* add stream.h

* fix ut

* fix cpu compile

* Use RWLock in GetAllocator

* Fix CI error
Co-authored-by: NChen Weihang <chenweihang@baidu.com>
Co-authored-by: Nzhiqiu <chenqiuliang@baidu.com>

* add completion_nofifier

* fix bug

* unregist event waiter

* add mul_lstm_fuse_pass ut

* update mul_lstm_fuse_pass ut

* update ut

* update ut

* update ut

* add CPU ut cmake setting

* update ut

* fix bug in pfp16

* fix hip

* fix hip

* add device-agnostic stream class

* add stream.h

* fix ut

* fix cpu compile

* Replaced pten::LoD with paddle::framework::LoD

* Overrided CPUVector with CUDAVector

* Refactored paddle::framework::Vector

* add conv+hard_sigmoid fuse pass ut

* update conv_elementwise_add_mkldnn_fuse_pass ut

* update conv_hard_sigmoid_mkldnn_fuse_pass ut

* update conv+hard_sigmoid and conv+hard_swish fuse pass ut

* update ut

* update ut

* First set of fixes

* - Make more likely to GetBlob find a blobs

* - Lint

add mem pool

* add align for WorkQueue

* add spinlock

* merge develop

* merge

* Add EventsWaiter

* Revert "Add EventsWaiter"

This reverts commit e206173aa9be7401b83a53581627bfaf557c8fb2.

* update EventsWater

* fix

* split workqueue files

* add more tests

* fix

* bugfix

* bugfix

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

* add mkldnn conv_elementwise_add_mkldnn_fuse_pass ut

* update mkldnn conv_elementwise_add_mkldnn_fuse_pass ut

* update conv_elementwise_add_mkldnn_fuse_pass ut

* update conv_elementwise_add_mkldnn_fuse_pass ut

* update conv_elementwise_add_mkldnn_fuse_pass ut

* restrict conv2d data_format in conv_elementwise_add_mkldnn_fuse_pass

* update conv_elementwise_add_mkldnn_fuse_pass OpCompat

* update conv_elementwise_add_mkldnn_fuse_pass ut

* update ut

* add new API: paddle.clone;Tensor.element_size;nn.utils.parameters_to_vector

* fix comment

* add flatten2_matmul squeeze2_matmul reshape2_matmul test case

* modify skip func to ignore_pass_case func

* rebuild CI

* add test_xx_matmul_fuse_pass timeout

* add test_map_xx_pass timeout

* add max_duration of test cast

* add trt skip

* add timeout

* del commented code

* add mkldnn reshape_transpose_matmul fuse pass ut and op version check

* update reshape_transpose_matmul_mkldnn_fuse_pass ut

* update ut

* update mkldnn batch_norm_activation fuse pass ut

* update ut

* update mkldnn batch_norm_act_fuse_pass ut

* update batch_norm_act_fuse_pass ut

* update ut

* CE fix

* more format

* update seqconv_eltadd_relu_fuse_pass ut

* update ut

* update ut

* update ut

* update squared_mat_sub_fuse_pass ut

* update ut

* update ut