[NewIR]Add NOT_FOR_INFER to prune Inference Library Size and Split VJP CodeGen into pd_op_vjp.cc (#56352)

* [NewIR]Prune Inference Library Size and Remove IR Dialect

* remove options

* add NOT_FOR_INFER

* fix pd_vjp.cc

* polish deps

* fix code style

* fix unittest

* fix cmake

* fix inference CI

This reverts commit 422a565d.

* fix openssl bug

* fix openssl bug

* fix openssl bug

* test

* upgrade openssl version

* fix

* fix

* update

* fix openssl bug

* fix

* Update ci_dockerfile.sh

---------
Co-authored-by: NYUNSHEN XIE <1084314248@qq.com>

* fix proxy problem

* fix leveldb bug

* fix proxy problem

* [CINN] Fix compile error

* delete useless env

* fix gflags conflict

* fix args used by ut

---------
Co-authored-by: NZzSean <18818272991@163.com>

* [cmake] add boost third_party cache

* del boost

* Update ginac.cmake

* Update ginac.cmake

* Update ginac.cmake

* add absl to third_party

* fix

* ci test

* ci test

* ci test

* rollback

* add isl third_party cache

* fix bug

* Update cinn.cmake

* onednn: rename macro to PADDLE_WITH_DNNL

* onednn: rename macro to CINN_WITH_DNNL

* fix xpu

* change download dir

* fix dir

* fix dir

* roll back

* bakup

* code style check

* fix wrong path

* roll back

* update paddle2onnx

* update codestyle

* myStupidBug STREQUAL->EQUAL

* Work for fa-2 padded fwd. Code to be cleaned.

* Work for fa2 unpadded fwd.

* Work for padded-bwd, dk get small diff on np.random.seed(0)

* Anyway I pass paddle's utest, except return softmax without dropout.

* Clean code.

* Modify interface.

* Clean code and add some check.

* Easy compile for dev.

* Fix ci.

* Fix ci-build.

* Add std c++17 option again.

* Limit max job when compiling fa2.

* Remove const_cast

* Add fwd params, to be cleaned.

* Clean code.

* Add bwd params.

* Clean code.

* Add enforce.

* Use v2.0.4

* Pass RNG state to fa2 capi

* Fix review.

* Add assert

* Skip compile for sm less than 80.

修复Paddle With CINN 执行 python单测下op_mapper 时报gflags符号undefined问题。

support stride

* [CINN Frontend] Optimize user interface, polish fuse_pass classes"

* Fix some compile error

* [cmake] add third party jitify cache

* fix bug

* fixed

* fix bug

* Clarify cinn/ir dirs [Part1]

* addition of cinn/ir/op dir

* change header inludsion of ir/ir_operator.h to ir/op/ir_operator.h

* merge with develop changes

* relocate libschedule_desc_proto.a

* remove extra ir_schedule_error.cc

* addition for schedule/ir_schedule_error files

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* remove rocksdb change

* fix

* fix

* fix

* fix lite

* fix

* Update lite.cmake

---------
Co-authored-by: Nrisemeup1 <62429225+risemeup1@users.noreply.github.com>

In XCCL 1.0.53.6 there is a bugfix to fork scenario in RDMA, which
was the root cause of dataloader segfault problem.

* fix

* fix

* fix

* fix

* Update CUDNN Frontend API to v0.9.1
- Remove old patches
- Remove workarounds that are no longer needed

* Fix test_switch_autotune

* new group fuse pass api

* fix header

* update

* change logic of get master node to fix bug

* revert update for ReduceFuseReduce

* modify according review

* modify by review

* refine

* update

* fix code-format

* squash pick the poc code
* fix build after rebase
* fix int8 conv and fc uts
* Fix and clean-up Get_SRC_Scale_Memory
* fix floating point fc uts
* fix test_analyzer_int8_googlenet
* test_analyzer_int8_mobilenetv1
* fix int8 mobilenet v2 and v3
* fix build error after rebase
* [oneDNN] rename library version
* fix conv bias datatype
* try to fix import error
* fix rebase error
* [oneDNN] pack library into python wheel
* add MKLDNN_SHARED_LIB_3 to env_dict
* fix test_analyzer_bert
* fix fill_constant op kernel
* fix ernie and matmul op ut
* fix softplus ut
* fix conv+relu6 fusion ut
* fix hardswish fusion
* fix quant+transpose fusion ut
* fixsgd ut
* fix int8 matmul with flatten
* fix fc+scale fusion
* fix conv/matmul+gelu fusion uts
* fix rebase error
* Revert "fix conv/matmul+gelu fusion uts"
This reverts commit 47eb5e49972bd8f7271a233def9bfb3e98ce78e1.
* upgrade to onednn v3.1
* remove older version onednn
* use densetensor::data() for achieving mean and var in layernorm impl
* comments for atol of integer tests
* fix clang-format
* Revert "remove older version onednn"
This reverts commit 783e57ddfd4401254596eae7d47adb9b03590c09.
* improve binary handle
* fix expand kernel
* Revert "use densetensor::data() for achieving mean and var in layernorm impl"
* always use forward_inference for conv
* remove activation scales
* rollback changes to mkldnn.cmake
* address comments
* port changes to dequantize kernel
* fix merge error
* fix fused_elementwise_kernel
* upgrade onednn version to v3.1.1
* fix some approval error
* fix error msg format
* remove old onednn libs
* try to fix symbolic link issue
* fix cinn test case segfault
* do not explicit link test with onednn
* remove unnecessary changes
* integrate CINN with onednn v3
* link with mkldnn project
* fix cinn build file

---------
Co-authored-by: NTomasz Socha <tomasz.socha@intel.com>
Co-authored-by: NChen, Xinyu1 <xinyu1.chen@intel.com>
Co-authored-by: Ntianshuo78520a <707759223@qq.com>

* fix WITH_SHARED_IR option type

* rename WITH_INFERENCE_NVTX to WITH_NVTX and fix compile bug

* update

* remove some pybind interface in cinn to fix compile problem

* modify cmake

* fix cmake

* add log for build cinn whl

* fix ninja for cinn

* fix conflict

Pcard-72384
paddle with cinn的情况下支持cinn的c++ ut。 统一了cinn和paddle的third_party的c++版本，cinn third party不使用C++17

This reverts commit 15c87528.

* set cmake_cuda_architecuture

* set cmake_cuda_architecutres

* set cmake_cuda_architecutres

* Add XPU plugin to support the customized ops or improve the performance of the fusion ops based on hand-written xpu micro kernels.

* refine README.md