* add wait_server_ready

* fix remove bug

* [NPU] add sigmoid and sigmoid_cross_entropy_with_logits

* [NPU] add debug logs

* [NPU] delete debug logs

* [NPU] reset to paddle.NPUPlace(0)

* [NPU] add more unittests for sigmoid_cross_entropy_with_logits

* delete useless c++ code, and delete check_dygraph=False

* [NPU] delete repeated codes of unittests

* [NPU] fix licence code style

* [NPU] add attrs (now do not support) check for sigmoid_cross_entropy_with_logits

As the title, this PR converts all blocks in program into SSA sub graphs and it is guarded by flag

* add clear_float_status op

* refine infershape

* fix typo

* refine check_finite_and_scale

* refine code

* Add NoNeedBufferVarsInferer

* fix code style

* fix bn

* fix

* add unittest

* fix cpu

This PR creates a class to process the program at the C++ level. Currently, this class has one class method:
GetInputsOutputsInBlock()

* Add DCU backend support for custom ops

* Added checks for DeviceCopy and renamed some macros

[pass enhance] make the attribute check only object to which defined in op proto. test=develop (#34146)

* Refine Constructor logic of ParallelExecutor

* Replace executor into ParallelExecutor in run_program_op

* add scatter-gather opt

* fix topo for pp

* rename function

* adam add input SkipUpdate

* add unittest

* add npu unittest

* fix xpu compile

* remove param stream

* Support sccache to speed up compilation on Windows

* Support sccache to speed up compilation on Windows

* Fix gather_op by adding OurOfRangeCheck for param[Index]

* Code Optimization

* added printing tensor's format

* added suggested changes

* test=cpu-gpu-py3

* test=cpu-gpu-py3;notest

* notest;test=cpu-gpu-py3