* add recorded cuda memory apis, fix typo, test=develop

* add more ut, test=develop

* follow comments, test=develop

* fix py35 incompatible issues, test=develop

* encapsulate the PaddleTensorToLoDTensor, test=develop

* serialize the pd_tensor, test=develop

* serialize tensors to file, test=develop

* add fluid.device_guard to specify the device type for Op

* update ScopeBufferedSSAGraphExecutor&AsyncSSAGraphExecutor&ThreadedSSAGraphExecutor&FastThreadedSSAGraphExecutor&ParallelSSAGraphExecutor&ParallelExecutor for fetching unmerged results.

* add the unit test for fetch_unmerged.

* update ut for multi-card and multi-cpu.

* add the error message and the user suggestion in FetchOpHandle. test=develop

* add lodtensor share memory & serialization, test=develop

* fix windows compile error, test=develop

* deal vartype pickle & fix unittest matching error message, test=develop

* update timeout variable name, test=develop

* refactor memory map implement, test=develop

* clear mmap file discripter when exit unexpectedly, test=develop

* remove the child process fd in advance, test=develop

* remove mmap fds after Queue.put in child process, test=develop

* add hard unittests for register exit func, test=develop

* fix python2 compatibility problem in unittest, test=develop

* fix exception unittest error, test=develop

* polish code based review comment, test=develop

* user can call dataset.set_download_cmd to set its customized download cmd
* add UT to cover this scenario

* Refine adam op, test=develop

* Fuse kernels together to reduce cpu time.

* Refine paddle enforce, test=develop

* Remove some comments, test=develop

* Refine code,test=develop

* Refine cuda kernel, test=develop

* Refine code according to comments, test=develop

* support cond in clone, test=develop

* refine code, test=develop

* refine code, test=develop

* follow comments, test=develop

* refine code, test=develop

* Add two types of Metric Calculator: MultiTaskCalculator & CmatchRankCalculator.
* Add a config for DynamicAdjustChannelNum function to denote whether we will discard the remaining instances when they are not be distributed evenly.
* Remove CPU code in Pull/PushSparse and we will add it back when testing it fully.
* Fix some known issues: such as copying persistable vars after one epoch running.

* add sync communicator and implement

* add python flag to control profile level test=develop

* add thread barrier for the compiled program

* update

* update test=develop

* update compile set test=develop

* update compile set test=develop

* update test=develop

* update test=develop

* update test=develop

* update compile setting test=develop

* update compile setting test=develop

* update run demo test=develop

* update test=develop

* update test=develop

* fix test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update format test=develop

* update format test=develop

* update style test=develop

* update style test=develop

* change style test=develop

* change style test=develop

* change style test=develop

* add dataset unittest test=develop

* update test=develop

* update for record test=develop

* udpate style for record test=develop

* update for record test=develop

* update for record test=develop

* update for record test=develop

* fix format test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

* update test=develop

Compile without nccl deps. [1/2]
Co-authored-by: N石晓伟 <39303645+Shixiaowei02@users.noreply.github.com>

* Add the first implememtation of fusion_group op #19621 (#3)

* Add the dynamic load of nvrtc, and support runtime compiling of CUDA kernel using nvrtc.
test=develop

* Call CUDA driver api to launch the kernel compiled by nvrtc.
test=develop

* Disable for mac and windows.
test=develop

* Refine the codes to support manually specified num_threads and workload_per_thread.
test=develop

* Refine the CUDA kernel to support large dims.
test=develop

* Add DeviceCodePool to manage all device codes.

* Add the first implementation fusion_group op.

* Add unit-test for fusion_group op.

* Add the check of result.

* Add the check of nvrtc in unit-test.
test=develop

* Add comment to explain the inputs, outputs and features of fusion_group op.
test=develop

* Disable fusion_group op for mac and windows.
test=develop

* Make the compiling of device code return status instead of hanging up.
test=develop

* Add the check of whether there is CUDA driver library, and do not core dump when failing to call the CUDA driver API.

* Unify fusion_group_op's input and output names.
test=develop

* Add the check of CUDA driver library in unittest.
test=develop

* Enable generating code for a given subgraph. #21126 (#4)

* Enable generating code for a given subgraph.

* Support sorting the subgraph.

* Remove the rearange of expressions because we use the sorted subgraph directly.

* Enable generating code for a subgraph which is composed of grad ops.

* Use expression information to check the accuracy in unittest.

* Separate load and store from computation expressions.
test=develop

* Improve the loading statements in generated codes.
test=develop

* Remove unused arguments from formal list.
test=develop

* Enable the detection of subgraph of grad ops.

* Generate code for detected subgraph in fusion_group_pass.

* Add an option in BuildStrategy to enable fusion_group_pass and add unittest.
test=develop

* Fix a bug when checking whether the shape of all inputs are the same.

* Add debug information.

* Remove subgraph_detector from inference/analysis to the common framework/ir directory. (#5)

test=develop

* Call subgraph_detector in fusion_group pass.
test=develop

* Disable fusion_group when WITH_GPU is OFF.
test=develop

* Refine all PADDLE_ENFORCE message.
test=develop

* Fix the case that some inputs are not defined in grad ops, and set op_role for fused op.
test=develop

* Follow review comments.
test=develop

cmake选项中添加了WITH_NCCL，显示指定是否编译NCCL的部分代码，WITH_NCCL默认打开，但如果WITH_GPU为OFF，则关闭WITH_NCCL

添加了PADDLE_WITH_NCCL定义

单机单卡能够关闭NCCL编译，多卡的话需要默认打开NCCL，如果关闭NCCL，则只能使用单卡
Co-authored-by: N石晓伟 <39303645+Shixiaowei02@users.noreply.github.com>

* add int16 support, test=develop

* add test, test=develop

* fix typo, test=develop

* fix dtype error in slice, test=develop

* add GeneralRoleMaker which is for general usage
* test=develop

remove unnecessary template.

* use function instead of lambda, test=develop

* follow comments, test=develop

* Implement a common python unittest to test the ir passes.
test=develop

* Save the results in np.array and support to startup on CPU.
test=develop

* Fix the unittest.
test=develop

* Add check_program to check whether the optimized program is different from the origin one.
test=develop

* Remove the inferface all_ops.
test=develop

* Add exception test in pass_test.
test=develop

* add half_async in the communicator
* fix DistributedStrategy

* add multiprocess for dygraph data loader, test=develop

* polish code & add safe gurad, test=develop

* refactor dygraph dataloader & add signal handler, test=develop

* fix member initializer compile error on ci, test=develop

* fix member initializer compile error one more, test=develop

* remove useless config, test=develop

* skip windows incompatible problem, test=develop

* add unittest for coverage, test=coverage

* add more exception unittest case, test=develop

* deal with signal handler coverage, test=develop

* polish code & add signal handler tests, test=develop

* deal with coverage ci problem, test=develop

* split data loader test & coverage ci fix, test=develop

* remove test_imperative_data_loader_with_exception, test=develop

* remove singal process except test case, test=develop

* add exception tests again & remove sample list test, test=develop

* split normal and exception unittests to diff class, test=develop

* polish doc for use_multiprocess effect in static mode, test=develop

* add collective communication library in fleet to replace mpi
* test=develop

* add bn and relu fuse pass

* add op attr assert and dtype assert

* fix some inputs&&outputs bugs for the fused op and pattern.

* add the unittest for fuse_bn_act_pass. test=develop

* use normative enforce statements. test=develop

* add the cpu test. test=develop

* add the support of batch_size=1 for the bn with relu op. test=develop

* add the error type for paddle throws. test=develop

* add fused_batch_norm_act and fused_batch_norm_act_grad to op_has_unsed_vars_white_list. test=develop

* modify fc to linear in sample code, test=develop

* remove FC, test=develop

* remove warnings, test=develop

* drop fluid/imperative/README.md , test=develop

* change fc to linear, test=develop

* polish code style, test=develop

* add distributed_strategy

* Fix openblas to support compile on Windows when WITH_MKL=OFF

* python zero copy inference
* support delete inference pass

* add some debug flags to auto growth allocator, test=develop

* add comments about auto growth, test=develop

The fixed bugs:

1. The condition sub-graph is not pruned
2. When backward graph is extremely simple, the whole backward ops are pruned.

* fix compiled error of butil when with_pslib=on and with_testing=on
* test=develop

* polish cmake, test=develop

* add current directory to LD_LIBRARY_PATH, test=develop

* add no_need_buffer_slots interface to pybind

* refine dygraph dataloader & polish related code, test=develop

* refine code based review comment, test=develop

* add op function generator, test=develop

* add unittest, test=develop

* follow comments, test=develop

* fix windows compilation problem, test=develop