Describe the current behavior / 问题描述 (Mandatory / 必填)

wide&deep网络在910环境8p训练，性能5374683/fps达不到5500000

Environment / 环境信息 (Mandatory / 必填)

• Hardware Environment(Ascend/GPU/CPU) / 硬件环境:

Please delete the backend not involved / 请删除不涉及的后端:
/device ascend

• Software Environment / 软件环境 (Mandatory / 必填):
  -- MindSpore version (e.g., 1.7.0.Bxxx) :r1.8 B010 commit_id:42306df4
  -- Python version (e.g., Python 3.7.5) :
  -- OS platform and distribution (e.g., Linux Ubuntu 16.04):
  -- GCC/Compiler version (if compiled from source):
  run包:HiAI/HISI_C82/20220518

• Excute Mode / 执行模式 (Mandatory / 必填)(PyNative/Graph):

Please delete the mode not involved / 请删除不涉及的模式:
/mode graph

Related testcase / 关联用例 (Mandatory / 必填)

test_ms_wide_deep_criteo_train_infer_910_8p_0001.py

Steps to reproduce the issue / 重现步骤 (Mandatory / 必填)

1. get code from models
2. sh run_multinpu_train.sh

Describe the expected behavior / 预期结果 (Mandatory / 必填)

网络训练成功，8p训练性能能达到5500000/fps

Related log / screenshot / 日志 / 截图 (Mandatory / 必填)

epoch time: 39788.940 ms, per step time: 123.568 ms
epoch time: 7635.088 ms, per step time: 23.711 ms
epoch time: 7643.314 ms, per step time: 23.737 ms
epoch time: 7627.080 ms, per step time: 23.687 ms
epoch time: 7617.906 ms, per step time: 23.658 ms
epoch time: 7673.134 ms, per step time: 23.830 ms
epoch time: 7691.547 ms, per step time: 23.887 ms
epoch time: 7678.079 ms, per step time: 23.845 ms
epoch time: 7710.388 ms, per step time: 23.945 ms
epoch time: 7671.512 ms, per step time: 23.825 ms
epoch time: 7676.350 ms, per step time: 23.840 ms
epoch time: 7684.426 ms, per step time: 23.865 ms
epoch time: 7702.860 ms, per step time: 23.922 ms
epoch time: 7678.710 ms, per step time: 23.847 ms
epoch time: 7668.595 ms, per step time: 23.816 ms

Special notes for this issue/备注 (Optional / 选填)

走给曹杰文