* support pinned, test=develop

* support async_write, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine,test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* refine, test=develop

* add processgroup_nccl

* add pten::Place data structure.

* update ci problem

* fix ci problem

* update

* using platform::Place=pten::Place

* remove BOOST_GET_CONST for CPUPlace and GPUPlace

* compile pass 25%.

* compile pass 45%

* compile pass 60%

* remove boost_get for xpu npu mlu and ipu

* compile pass on cpu and gpu.

* fix compile problem

* fix compile error.

* update

* fix ci problem

* update

* ci approve

* fix ci problem

* fix ci eager test problem

* remove BOOST_GET_CONST

* fix npu compile

* refine structure for cuda and rocm

* update

* update

* update

* update

* unify calling cudaSetDevice

* fix compile

This PR is follow up of #28213. On that PR we tried to decrease GPU usage, however the CI still randomly failed. So I added retry logic for the initialization of nccl and cusolver. If the initialization failed, we can retry to avoid the random failure.

* Refine PADDLE_ENFORCE in paddle/fluid/platform
test=develop

* refine PADDLE_ENFORCE
test=develop

* add new macro BOOST_GET_SAFELY & unittests, test=develop

* add different macro type, test=develop

* fix get macro type in executor, test=develop

* four macro part change backup

* using one macro for all case, test=develop

* revert attribute change, test=develop

* change to three func to solve gcc4.8 bug, test=develop

* polish some details, test=develop

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

* fix warnings oof gcc 8 compilation, test=develop

* fix boost::bad_get, test=develop

* refine PADDLE_ENFORCE, test=develop

test=develop

* wip allreduce in op

* wip

* wip

* wip

* wip adding test

* wip for conflict with mp mode

* fix tests test=develop

* fix cpu build test=develop

* fix travis clang format test=develop

* fix cpu build test=develop

* update api.spec test=develop

* delete comment test=develop

* fix cpplint test=develop

* fix test=develop

* follow comment test=develop

* add file test=develop

* fix build test=develop

* update test=develop

* to be compatible with sync_bn, and fix mp mode in develop test=develop

* Support Sync Batch Norm.
* Note, do not enable it in one device.

Usage:

build_strategy = fluid.BuildStrategy()
build_strategy.sync_batch_norm = True
binary = fluid.compiler.CompiledProgram(tp).with_data_parallel(
        loss_name=loss_mean.name,
        build_strategy=build_strategy)

test=develop

* wip

* wip

* wip

* wip for test

* add fp16 tests test=develop

* fix cpu build test=develop

* fix test=develop

* fix py3 tests test=develop

* fix lr_scheduler dtype test=develop

* fix test=dvelop

* test fix ci compile test=develop

* fix build and merge test=develop

* fallback momentumop change to general test=develop

* make fp16 lr schedule simple test=develop

* fix ut test=develop

* fix tests test=develop

* remove fp16 learning rate cast test=develop

This reverts commit 3d750f9c.

* wip

* wip

* wip

* wip for test

* add fp16 tests test=develop

* fix cpu build test=develop

* fix test=develop

* fix py3 tests test=develop

* fix lr_scheduler dtype test=develop

* fix test=dvelop

* test fix ci compile test=develop

* fix build and merge test=develop

* fallback momentumop change to general test=develop

test=develop

* wip multi process multi gpu dist training

* workable for p2p

* update test=develop

* change back env name test=develop

* fix alloc init

* fix cpu build test=devlop

* fix mac tests test=develop

* refine code

* refine test=develop

test=develop

Fix code to support cpplint syntax check

test=develop

* support test using executor without reader

* run imgnet

* update fluid benchmark

* wip

* update

* update all models

* support pyreader

* update

* clean up

* make profile batches contollable

* update API.spec

* update scripts

* clean dockerfile

* update

* clean comments

* add scope argument docstring

* use num_trainers to determine nccl init comms

* 1. Create buddy allocator in each places before NcclBcast the variables
2. Check the memory usage of ALL gpus rather than the first one

* 1. Make NCCLGroupGuard guards only the ncclBcast part, which avoid ncclGroupEnd blocking the exception throwing
2. NOTE the usage of NCCLGroupGuard

* Remove the memory usage check of gpus

* Fix code style