* support default use_gpudnn=True

* fully support cudnn in phi

* add header file

* add white_list, verify accuracy

* phi support all cudnn

* opt affine_grad

* try different arches of pretrained_model

* try different arches of pretrained_model

* add debug string

* debug eager_method

* add debug string, pass all local ctest

* polish all debug code

* delete use_cudnn relevant code autogen

* fix depthwise_conv2d

* Share all other members of Tensor except use_cudnn

* polish codes according to review opinion

* polish codes according to review opinion, fix bug

* polish codes according to review opinion, opt performance

* polish codes according to review opinion, fix pooling.py

* do not calc reduce_all in eager mode

* refine python c cast list

* refine

* refine

* refine

* refine

* refine

* refine

* refine

* refine

* refine

* dense tensor in eager mode support data_ptr

* add nparray case for basic operator

* fix unit test

* fix unit test

* add unit test

* fix unit test

* fix doc of NPUPlace

* fix doc of NPUPlace, test=document_fix

* refactor: replace Collective & PointToPoint with NCCLEnv

* refactor: rename to RunFnInNCCLEnv

* refactor: pass std::function by value

* fix bug of zero_allocator in host

* fix test compile bug

* add unittest

* update test

* fix device id issue for xpu eager

xpu device id is not correctly set in eager mode, thus vars are on dev0 unless
XPUDeviceGurad is called, leading to this error message for all node rank != 0:
"NotImplementedError: (Unimplemented) Place Place(xpu:0) is not supported."

* fix typo

* fix pybind error

Support bfloat16 for adamw and adam optimizer. Fit the lr for pure bf16 training with tensor fusion. (#48041)

* add bfloat16 for adamw

* set lr not to bfloat16 for pure bf16 training

* update the logic

* update the adamw optimizer

* support bfloat for adam

* refactor: simplify send, recv interfaces

* refactor: rm send_partial, recv_partial, all_gather_partial

Get grads types from cpp for adam to speed up

* standard api

* fix sparse bugs

* fix xpu bugs, test=kunlun

* remove hard code for custom unittest

* open ci, test=kunlun

* deal with conflict

* XPU support eager mode

* add unittest for XPU eager mode

* minor bugfix

* minor bugfix, test=kunlun

* correct copyright info

* 1. remove unsed vars/funcs
2. ProcessGroupBKCL inherit from ProcessGroupStream

* bugfix for fp16 in eager mode multi-card, test=kunlun

* rebase & fix a few issues

* use new processgroup interface, test=kunlun

* fix compile issue, test=kunlun

* refactor: send, recv, send_partial, recv_partial

* refactor: rm useless const ref

* Get params and grads in cpp to avoid gpu idel time

* Using python param instead of cpp return param to fix test_asp_optimize_dynamic.py

* Get grads from cpp and construct params_grads on python

* Check meta and remove comments

* Analysis API interface for disabling fc passes

* Unit tests corrected

* Python API added

* test runs only when PADDLE_WITH_MKLDNN

* Fc op changed to relu in matmul_op_test

* Disable fc passes in tests where acc drops

* code formating

* Unit test for analysisConf added

* Unit test gpu added

* fc passes disabled when iterations=0 in gru test

* style

* passes disabled when fp32 in gru test

* fc passes disabled in lstm test

* Import from inference, not fluid in doc

* refine python call error report

* Get three grad lists in CPP to avoid gpu idle time

* Support legacy mode

* fix cc_library link python lib

* [Auto Parallel] Improve the c++ dist attr

* [Auto Parallel] Modify test_program.py

* Support custom stream for standalone executor
Co-authored-by: NYulong Ao <aoyulong@baidu.com>

* [Auto Parallel] Improve the c++ dist attr

* [Auto Parallel] Modify test_program.py

* [Auto Parallel] Add the missiong import