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

* add vectorized bfloat16 atomicAdd

* fix compile error

* fix compile error again

* fix V100 compile error

* fix V100 compile again

* add_fill_any_like

* add_fill_any_like

* elementwise_op

* add teller

* modify ut

* comments

* modify ut

* return

* modify

* Enable bf16 in oneDNN bilinear_interp kernel

* Fix bilinear_interp_v2 not enabled in models

* Remove unnecessary checks

* remove avx check

* fix bug;

* refactor: update pg custom

* fix: use new api in ut

* fix: typo

* revert: recover legacy apis

* fix: add GetDeviceContext

* optimization for ln

* fix

* added output to gpd

* added formatting

* fix

* cleanup unused code

* unify is_int8 is_bfloat16

* Simplify matmul_v2 FWD kernel

* remove RunKernel methods

* remove import namespace

* remove headers

* clean fluid/phi cross imports

* remove fluid axpy_handler

* delete fluid methods

* activations

* OneDNNMemDesc

* MKLDNNFormatForSize

* MatchShapeToLayout

* MKLDNNMemoryFormat

* MKLDNNFormat

* ReorderMKLDNNHandler

* to_void_cast

* review suggestions

* interpolate

* remove fluid depedency

* refactor: simplify send, recv interfaces

* refactor: rm send_partial, recv_partial, all_gather_partial

* feat(ipu): add model_runtime backend support in IPU.

* fix(ipu_executor): fix error message format.

* fix(ipu_executor): fix format.

* fix(ipu_executor): fix format again.

* fix(ipu_executor): fix format again.

* fix(ipu_executor): fix format again.

* refine shape op in new_exe

* Revert "refine shape op in new_exe"

This reverts commit 0e0336ddc5eede3da019b348a0bcc0ef0f3be64e.

* refine shape op in new_exe

* refine shape expected_kernel_type

* add SelectedRows check for shape op

* refine code

phi::Alloc() complains about missing device_allocator_

* generate static graph code for some ops by yaml

* remove deleted files

* update cmake

* update cmake

* udpate cmake

* migrate prelu

* remove cache

* review fixes

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

* add roformer pass&&plugin（novarlen）

* 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

* skip_merge_layernorm

* add UT

* modify comments

* add ci check for code-gen script

* update