Merge pull request !5111 from YuJianfeng/int64

Merge pull request !5315 from limingqi107/master

Merge pull request !5171 from liangzelang/fix_global_step_error

Merge pull request !5055 from lirongzhen1/int64

Merge pull request !5140 from zyli2020/add_cuda_path_check

Merge pull request !5035 from zhoufeng/xiu-ba-ge

Signed-off-by: zhoufeng <zhoufeng54@huawei.com>

Merge pull request !5020 from lianliguang/master

Merge pull request !3775 from lianliguang/test-xiu-bug

Merge pull request !4934 from zyli2020/bug_fix

This reverts commit e2b346d5af.

set parameter device dtype using it's infer dtype && set transdata's abstract

Merge pull request !4830 from yuchaojie/gpu_allreduce

Merge pull request !4472 from anthonyaje/gpu_queue_size

Merge pull request !4843 from xychow/remove-ccsrc-common-h

replace frontend/operator/ops.h in backend with base/core_ops.h as
backend should not use any frontend-only primitive

Merge pull request !4557 from caifubi/profiling

Merge pull request !4731 from gukecai/parallel-genmask

Merge pull request !4684 from 治愈系潇洒哥/master

Merge pull request !4392 from xychow/use-float16-in-arm-neon

Merge pull request !4428 from AdelShafiei/opoverflow2

Merge pull request !4656 from jjfeing/master

Signed-off-by: zhoufeng <zhoufeng54@huawei.com>

Merge pull request !4652 from lvchangquan/transdata

Merge pull request !4585 from baihuawei/broadcast

Merge pull request !4609 from chenzhongming/new_master

Other Authors: Harshvardhan Gupta, Li Chen

This reverts commit adb6ff6c78.