* Update the content of `test_parallel_executor_run_cinn.py`.

* Fix some bugs in the topological sort and `CreateNewSubGraph`.

* Update the CINN commit id used by Paddle.

* Update the unit test to `add+relu`.

* Update according to reviewers' suggestion.

* add TypeAdapter method for npu_op_runner

* add int64 supporting for elementwise_mul and reduce_sum

* add int64 supporting and UT for expand_v2, scale and reduce_max

* fix bug

* Revert "Align CTC grad scale same with ESPNet (#34729)"

This reverts commit 10f9644c.

* ctc grad compute on gpu

* add trainer desc config to distributed strategy

* code style modified

* data_feed set lod

* fix bug

* code style

* fix bug

* save load

* save load

* save unittest

* add unittest of the_one_ps

* unittest

* add todo in communicator sendsparse

* Fix bug when pre_layer_norm is false.

* GeneratePass support attr condition and mapping, test=develop

* fix coverage, test=develop

* add paddle.version.cuda and paddle.version.cudnn API

* fix little bug

* fix bug

* add doc string

* fix mkdir error

* fix windows path

* fix new paddle/version path

* fix unittest

* fix format

* add dcnv2 plugin

* revise completion for backward

* revise completion for update

* revise completion for update

* update unitest

* Add WIP version of elementwise_div_mkldnn without working dy grad

* Add dy gradient calculation implementation, disable broadcast tests

* Readd removed tests from static_mode_white_list

* Add bfloat16 gradient tests, remove int8 and uint8 support

* - Change the way dy grad is calculated to improve performance
- Refactor BinaryMKLDNNHandler to use a default parameter

* Change copyright year

* Refactor as suggested

* Attempt to bypass CI Approval
not accepting max_relative_error

* Fix formatting issue

本PR是fused_transformer的layer层代码，包含FusedFeedForward的layer层代码和FusedTransformerEncoderLayer的代码。

* fix fftshift/ifftshift on static mode
* update roll_op version
* add more test cases for fftshift/ifftshift

* add eigvalsh with is_test

* add eigvalsh op

* fix backward bug

* forward and backward, float and complex, unittest

* remove eigvalsh_helper.h

* remove changes of cusolver.h

* fix unittest

* fix unittest bug

* update code following eigh

* fix test

* update lapack

* pull develop

* update funcor

* fix unittest bug

* fix details

* add tensor_method_func

* fix notes

功能：本PR的目标是提高attention模块的计算性能。
为了减少框架层对op的调度开销，本PR通过在C++层手动实现attention模块，对外提供attention 大op；
为了减少防存开销，本PR采取了两种优化方法：
（1）在q,k,v计算时通过共享输入X，将该处的gemm，transpose和bias add从三次调用减少为一次；
（2）使用kernel融合优化技术，在不同cuda kernel之间通过寄存器传输数据；

* cache exception in child thread

* add ut

* fix ut

将 #35905 和 #35843 PR中新增的的python api接口移到incubate目录下。

* [NPU] fix argsort op, test=develop

* remove debug files, test=develop

* fix typo, test=develop

* address review comments, test=develop

* fix wrong trt dim when input dim is 2

* update leaky_relu and instance_norm converter unit test

* add instance_norm input dim check

* move signal apis

* move fft.py and signal.py to paddle/, fix typos

* fix relative imports from fft.py and signal.py

* fix typos

[Paddle-Inference]Add MatmulV2ToMatmul convert Pass, fix (matmul_v2, matmul, mul) convert pass, fix (matmul, mul) op_teller (#36652)

* new_Matmul2ToMatmulToMul

* new_Matmul2ToMatmulToMul

* fix paddle_pass_builder

* fix paddle_pass_builder

* fix paddle_pass_builder

* tem

* tem

* Add MatmulV2ToMatmul convert Pass; MatmulV2ToMul convert Pass

* Add MatmulV2ToMatmul convert Pass; MatmulV2ToMul convert Pass

* add matmul_broadcast_unitest

* fix op_teller

* In cpu parallel using gloo, add various length support for SelectedRows

* fix bug

* fix bugs

* fix by code review

* remove timeout

* fix pool2d convert case

* add pool2d convert test case for trt6

* [NPU] modifications for model ernie-1.0

* rollback 503003 and change cast to dtype

这个PR是fused_feedforward反向的代码

相关kernel实现：fused_dropout_act_bias, fused_residual_dropout_bias, fused_layernorm_residual_dropout_bias

fused_feedforward是一个融合算子，该算子对transformer模型的feed forward层的算子进行融合和封装，使得前端只呈现一个接口，通过融合减少部分访存和kernel launch的时间，以此提升性能。

* Add bincount op

* upload cpu version

* fix unitest

* fix unittest

* fix unittest

* fix en doc

* add more test

* fix en doc

* add more test case

* fix test

* fix input vailidation

* fix input check

* fix unittest

* fix test

* fix en doc

* Init the functions of CinnCompiler.

* Add the unit test for CinnCompiler.

* Fix some compilation errors.

* Update the UT of cinn_compiler.

* Use Decomposer&OpFusion passes in CinnCompiler::CompileGraph.

* Update some comments.

* Uncomment some includes in build_cinn_pass.cc.

* Use refs instead of ptrs as returned types of FindGraph & Compile in
CinnCompiler.

* Use the merged CinnGraphSymbolization functions in CinnCompiler.

* add some ops to train ssd on kunlun

* add some ops to train ssd on kunlun

* add some ops to train ssd on kunlun

* update cast op unittest

* update cast op unittest

* update cast op unittest

* update xpu cmake

* update cast unittest

这个PR只包含fused_feedforward前向的代码。

相关kernel实现：fused_dropout_act_bias, fused_residual_dropout_bias, fused_layernorm_residual_dropout_bias

fused_feedforward是一个融合算子，该算子对transformer模型的feed forward层的算子进行融合和封装，使得前端只呈现一个接口，通过融合减少部分访存和kernel launch的时间，以此提升性能。

* slice

* add UT

功能：本PR的目标是提高attention模块的计算性能。
为了减少框架层对op的调度开销，本PR通过在C++层手动实现attention模块，对外提供attention 大op；
为了减少防存开销，本PR采取了两种优化方法：
（1）在q,k,v计算时通过共享输入X，将该处的gemm，transpose和bias add从三次调用减少为一次；
（2）使用kernel融合优化技术，在不同cuda kernel之间通过寄存器传输数据；