* fix dtype cast in amp.

* add test case and update docs.

* remove set_prim.

* add bf16 support and bf16/fp16 unittest for pool2d

* add include files

* dbg

* reformat

* reformat

* modify code according to review comment

* remove duplicate code

* remove dup code

* remove useless include

* dbg

* move python/paddle/fluid/tests/unittests/xpu to test/xpu

* update CMakeLists.txt

* remove xpu in fluid/tests/unittests/

* add path to op_test_xpu

* fix incorrect path

* update test script

* fix test_adadelta_op_xpu error

* [AMP OP&Test] support bf16 for batchnorm

* codestyle

* Update batch_norm_grad_kernel.cu

* Update batch_norm_kernel.cu

* fix codestyle

* fix

* fix

* fix

* fix

* fix

* Update batch_norm_kernel.cc

* [Do NOT merge] Expr PR on Composite

* Expr PR on Composite

* Revert some compsite experiment

* Remove unnecessary composite code

* Add rsqrt as sub primitives

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* isamp

* gpu

* cpu

* noamp

* fix instance_norm

* fix

* fix unit test

* fix unit test

* add unit test

* fix

* add big data tests

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* add test case

* fix

* fix

* fix

* fix

* fix

* remove amp test

---------
Co-authored-by: Nheyanru01 <429520051@qq.com>

* 【Hackathon 4th No.13】为 Paddle 新增 Bernoulli API

* [Change]change unittest_py scipy version

* [Change]修改BernoulliNumpy的类型参数;优化静态图测试流程

* [Change]优化类的初始化及逻辑;增加0D相关测试用例

* [Test MV] fft

* Update test_spectral_op.py

* stack fp16 & bf16 support

* unstack fp16 support

* unstack bf16 support

* append stack fp16 ut

* add unstack

* recover unstack cpu kernel

* fix some issue for unstack ut

* delete unuse var

* add check_place

* fix inference err

* add bfp16 test for isfinite

* fixed for ci

* deal with comments

* fixed test

* skip test in cpu

* deal with comments

* fixed for ci

* fixed testcase

* fixed for ci

* fixed for testcase

* Add backend for to_static API

* [CINN] Reopen some prim+cinn unittest

* revert scatter for ci

* change cinn test timeout from 120 to 150

* add softmax_with_cross_entropybf16 test

* correct defalut value in testBF16/FP16 op

* fix test checkout/grad, add skipif

* autogen segment_pool

* delete legacy_dygraph about segment_pool

* remove legacy profiler

* rm test_parallel_executor_profiler

* add_fp_bf_for_flip_gaussian_random

* forget convert uint

* fix some error

* fix some error

* unique id for mesh

* rng ctrl

* support dropout

* register op

* adopt for recompute

* update unitest

* support pp

* add PoissonNLLLoss API

* update unittests

* Fix poisson_nll_loss init and update data type support

* remove type comment

* Update doc string

* Fix doc string erro

* Fix doc string math equation format

* Add float16 and bfloat16 support

* support set master_grad

* move register_hook to auto_cast

* update unittest

* fix fp16 test

* update for review comments

* Support two inputs of multihead attention named qk_multihead

* [Opt Performance] Optimize custom operator performance, reconstruct python API auto-gen, add cache and use const inference

* opt AutoGradMeta implementation

* remove profiler codes

* fix unit test

* change year, 2021->2023

* fix int64_t parse bug

* autogen softmax_with_cross_entropy

* fix error in softmax_with_cross_entropy version

* add strategy force_sequential_run

* remove flag

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* adjust defalut tolerance of output and grad

* fix a bug in the grad of OpTest

* fix the type of setting defalut value in optest, both forward and
backward

* add defalut

* fix test_sum_op

* adjust tolerance

* fix the tolerance of eager

* add bf16 and fp16 to the activation tests

* remove some fixs

* fix activation

* fix fp16

* fix gelu

* fix the activation tests

* add bfloat16 specialization to singrad and cosgrad

* fix bugs

* fix bugs

* add unittest

* add skip

* add fp/bf to rrelu/rrelu_grad

* git add rrelu

* fix bugs