* lazy creating work queue

* fix dry_run

* Use all sitepackages path as the library/include path (#42940)

* Fix several unit tests and increase the unit tests stability (#43670)

* Reduce gather op unit tests size and increase the timeout

* Add NVIDIA_TF32_OVERRIDE for multi-processes environment

* Remove record test for device event ut

* Fix 3 unittest errors (#43532)

* Fix test_fuse_resnet_unit failure

* Fix test_imperative_auto_mixed_precision failure

* Fix sparse_attention_op error

* Fix sparse_attention_op error

* Use fixed random seed (#43659)

* for CI test_collective_sendrecv_api
Co-authored-by: Nzlsh80826 <rewang@nvidia.com>
Co-authored-by: NShijie <505749828@qq.com>

* cherry pick form develop 43621

* code format

* paddle2onnx update to 0.9.8

* Correct elementwise quantization (#43693)

* [Bug fix] Do not quantize weights Y when matmul X and Y both other ops outputs (#43297)

* fix some matmul that X and Y both other ops outputs, do not dequantize the Y.

* fix CI format

* fix according to review
Co-authored-by: Njoanna.wozna.intel <joanna.wozna@intel.com>

* add bilinear_interp_v2 converter

* update op_teller.cc

* add unittest for bilinear_interp_v2 converter

* code format

* bug fix

* code format and add unittest

* remove merged modify in op_teller.cc

* code format

* code format

* fix scale init error

cherry-pick #42750。

QA反馈，#42750 优化后，solov2模型性能可提升6%，故cherry-pick到2.3。因#41096 将linspace python实现从fluid.layers.tensor挪到了paddle.tensor.creation下，该pr不在release/2.3分支中，故将#42750 中python修改同步到fluid.layers.tensor.linspace中。

[cherry pick] Support optional residual add in fused ops and slice large tensor for cudnn_softmax (#43719)

 [cherry pick] Support optional residual add in fused ops and slice large tensor for cudnn_softmax

cherry-pick #43635 #43681 #43474

    Only cherry pick format tool(clang-format, yapf, cmake-format) upgrade to release/2.3, lint tool such as cpplint will not move, because we are not going to fix cpplint error in release/2.3
    pre_commit.sh also is moved to release/2.3 so that both PR-CI-pre-commit and PR-CI-pre-commit-23 can works.
    pre install clang-format to avoid repeat installation due to pre-commit's multi-thread running.

* cherry pick #43088 #40664

* fix clang format

* Update CUDA and TensorRT version for CI

* disable ut

* Update TensorRT for CUDA 10.2

删除 layout autotune 中的多余打印
背景 ：layout autotune log会导致模型打印信息增多

* cherry pick from #43397

* fix code

* Rename dropout is test (#43098)

* replace dropout_is_test with is_test.
* improve atol on a100.

* fused_attention fused_feedforward api support Model Tensor Parallel (#42985)

* fix is_test bug in fused_feedforward. (#43508)
Co-authored-by: NLi Min <11663212+limin2021@users.noreply.github.com>

* fix bug of strided_slice (#43388)

* fix stride_slice bug

* fix bug

* fix bug of infer shape for slice (#43443)

* [EinsumOp] Polish forward logic and backward logic for optimize (#42603)

* change logic for optimize

* modifty

* merge

* change einsum_v2 as default and add new flags: FLAG_einsum_opt=1|0 (#43010)

* [EinsumOp] Make EinsumOp support bfloat16. (#43085)

* change einsum_v2 as default and add new flags: FLAG_einsum_opt=1|0

* make EInsumOP support bf16

* add unittest for BF16

* add condition for test_BF16

* fix bugs

* fix

* change the backward api to fit einsum op

Reduce amax/amin frobenius_norm_kerne原始实现为Eigen实现，文件编译时间较长，因此本PR将其替换为KP实现
删除DefaultElementwiseOperator中重复功能支持，减少elementwise_double_grad OP编译时间

解决onnxruntime后端依赖的protobuf跟框架或外部protobuf版本冲突问题

删除Broadcast function中rank例化以及Elementwise调用，降低编译时间。
从develop分支中的#42645 PR修改而来，由于develop分支与release分支相差较大，无法实现cherry-pick，因此针对release2.3重新提交PR.
Broadcast中关于rank的例化会导致底层模板展开较多，造成reduce_sum_grad_kernel.cu.o文件体积过大，修改后可以降低.o体积及编译时间

* Fix cond_block_grad error when handle no need grad vras

* Add comment and UT

* fix_multi_int8 (#42977)

* cherry-pick fix_multihead_int8

cherry-pick PR #42700

* pdnode_compare

* panode compare

* pdnode_compare

* [MLU] add mlu new profiler (#41138)

* [MLU] add mlu new profiler

* fix format

* [MLU] support add callback to stream (#41831)

* [MLU] add gather mlu kernel (#41969)

* [MLU] add mlu activation kernels (#41751)

set attr ignoreIndex type to string for custom_nllloss_op

部分 cheery-pick of #42534

    add class NameScopeHelper for adding namescope info
    添加更多 种类优化器状态的映射
    为 IpuStrategy 添加 compilation_progress_logger option 用于输出 编译进度
    部分代码清理和杂项优化

Reduce the number of threads per block of deformable_psroi_pooling to solve the bug where too many resources requested for launch (PaddlePaddle#42531) (#42533)

* Reduce time variation for cuda_managed_memory_test (#42458)

* Disable standalone executor for test_tensordot (#42476)

* enable graph-engine to return all id (#42319)

* enable graph-engine to return all id

* change vector's dimension

* change vector's dimension

* enlarge returned ids dimensions

* change sample result's structure to fit training (#42426)

* enable graph-engine to return all id

* change vector's dimension

* change vector's dimension

* enlarge returned ids dimensions

* add actual_val

* change vlog

* fix bug

* bug fix

* bug fix

* fix display test

* singleton of gpu_graph_wrapper

* change sample result's structure to fit training

* recover sample code

* fix

* secondary sample

* add graph partition

* fix pybind
Co-authored-by: NDesmonDay <908660116@qq.com>
Co-authored-by: NDesmonDay <908660116@qq.com>

* fix paddle-ort python bug

* fix paddle-ort python bug