* Add quick gelu and fused bias add kernel

* fix annotation

* remove useless code

* add fast gelu option and set it in multi transformer op

* add flag to restrict if use fast gelu approximate

* fix flags conflict

* fix use tanh function instead

* add cudart version limit

* use phi fast tanh func

* fix comment

* fix typo `Fasle`/`Flase` -> `Flase`

* fix typo `Ture` -> `True`

* remove needless using tensor

* remove needless using tensor

* resolve conflict

* replace tensor using

* fix format error

* revert needless changing

* fix rocm and npu compile error

* fix cinn compile error

* fix format error

* fix mkldnn format error

* fix mkldnn format error

* fix cinn compile error

* fix cinn compile error

* fix cinn compile error

* resolve conflict

* remove cudaDeviceContext

* remove more template

* fix rocm compile

* remove alias name CUDADeviceContext

* fix compile

* fix tests

* revert changes

* Support optional residual add in fused_attention and fused_feedforward.

* Add checkpoint and add the check of add_residual when pre_layer_norm is false.

* Add TODO and change the python api to add add_residual argument.

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

* rename pten dir to phi

* rename namespace to phi

* rename infrt pten dir to phi

* resolve conflict

* rename pten to phi in cmake

* revert all infrt change

* change needed files

* fix infrt failed

* fix inference failed

* move paddle/operators/math/functors.h
* move paddle/operators/math/compound_functors.h

* Add fp16 support for scale/bias for fused_layernnorm_residual_dropout_bias op.

* Remove useless code.

* Remove useless code.

* Optimize layer_norm fwd when cols is 1024.

* Remove useless code.

* Minors.

* Minors.

* Modifications accordding to reviews.

* Minors.

* Optimize layer_norm bwd kernel when cols is 1024.

* Polish layer_norm_bwd_1024 kernel.

* Limit ln_bwd_1024_kernel to paddle_with_cuda.

* Fix double type compile error.

* Add optimization of ln bwd for fused_dropout_add_ln op.

* Polish codes.

* 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

* Add fp16 support for scale/bias for fused_layernnorm_residual_dropout_bias op.

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；