* support some prim ops for bf16 dtype

* remove useless code

* Add outputs to check_numerics_kernel.

* Add check_numerics to yaml.

* Add API and unittest.

* Add check_nan_inf_level as argument of check_numerics_kernel.

* Add more unittests.

* Fix static API implementation and unittest.

* Move the implementation of check_numerics to paddle.amp.

* Fix import error.

This reverts commit dd99a115.

* add vector type support for program translator

* polish

* support basic attribute type

* resolve conflicts

* add verify for combine/slice and unittests

* polish

* support more type in attribute translator

* modify by reviews

* fix merge mistakes

* refine code

* refine code

* add interface

* fix: op name normalization

* fix typo

* refactor input translator

* fix merge conflicts

* fix op normalizer bug

* refactor attribute translator

* fix bug

* refactor output translator

* fix typo

* fix

* fix approval error

* fix coverage

* fix op_compat parser

* fix merge conflicts

* fix merge conflicts

* fix merge conflicts

* fix merge conflicts

* fix merge conflicts

* refactor scalar attribute

* draft

* fix

* fix op build

* fix op build

* temporarily save

* adpat mutable attribute

* refine op_comat_gen process

* fix merge conflicts

* fix merge conflicts

* fix merge conflicts

* complete dialect attribute printer and refine ir_throw

* polish code

---------
Co-authored-by: Nzhangbo9674 <zhangbo54@baidu.com>

* [CustomDevice] add sarding support

* update

* add submodule pocketfft

* add submodule xbyak

* remove shallow_copy arg

* [cmake] add cusparselt third_party cache

* fix

* fix ci

* fix ci

* unary bitwise_not adapter tensorRT8.6 in Paddle-TensorRT

* one_hot_op ci fix

* Update test_trt_inspector.py

delete log

* replace requests with httpx

* set timeout=3

* replace requests with httpx

* replace request with httpx

* test

* repalce requests with httpx

* test

* replace requests with httpx

* replace requests with httpx

* modify paddle_build.sh

* fix bug

* [generator][cmake] add pip install check

* [generator] clean OUTPUT_VARIABLE

* set tanh_mutilply_double comp api as default

* modify timeout time

* rollback mutiply

* rollback mutiply

* [CINN] reopen mean/cumsum/instance_norm op's prim+CINN test

* remove repeat test_mean_op in cmake

* refactor scalar attribute

* fix op build

* fix merge conflicts

* fix coverage ci

* add ldexp api

* fix ldexp

* Update math.py

fix math.py

* rewrite the ldexp

* Simplify ldexp implementation

* fix codestyle

* Update math.py

* Update math.py

* modify the test_ldexp.py

* fix the input bug of np.ldexp

* fix the bug of np.ldexp in windows

* modify the ldexp function and add the dtype check of output

* Update test_ldexp.py

* fix the dtype

* fix codestyle

* Update python/paddle/tensor/math.py

add note for code description
Co-authored-by: Nzachary sun <70642955+sunzhongkai588@users.noreply.github.com>

---------
Co-authored-by: Nzachary sun <70642955+sunzhongkai588@users.noreply.github.com>

* fix_bug

* fix bug

* update roll convert

* [AutoParallel] fix sharding for 0D tensor and amp-o1

* add amp for sharding unittest

* Fix compilation error by using thrust

* fix

* update

* update

* update

* update

* update tag

* fix cutlass

* update cub tag

* update

* update

* add pslib cache

* add pslib_brpc cache

* update PS_LIB_SUBDIR var name

* [0D-Tensor] Support unary op, fix test_activation_op

* resolve conflict

* restore TestGelu

* polish codes according to comments