testcase

bug caused by lazy malloc of tensor memory. while batchsize large than
1, output tensor is divided(shared memory), and each divided tensor do malloc.

destroy base op after fuse

feat(subgraph): execute cnml compute forward asynchronously

outputs wrongly

(ref): add sigmoid and elementwise_mul x86 kernel interface

    1. disable conv activation pass by default
    2. set fc_fuser'param with_relu false while mlu fc kernel does not
    support relu
    3. change fc filter shape from 2 dim to 4 dim while input dim == 4
    4. add ToFile func in mlu tensor for debug convenience
    5. enable 4-dim input in elementwise_ops
    6. add transpose2d in utility.cc

           remove current logic for mlu

Develop set input layout

target -> subgraph -> target -> precision -> layout

Refactor input nchw layout

test=develop

* [LITE][OPENCL] enhance elemul, fix reshape for opencl. test=develop

* comment loop test for pad2 opencl kernel. test=develop

* [LITE][PROFILE] add sequnce statics for precision profiler. test=develop

* rename elementwise_sub_grad, test=develop

* rename Grad to GRAD, test=develop

* deal with case that Y@GRAD is empty, test=develop

* fix ci for test_mobilenetv1/v2. test=develop

* [OPENCL][CI] add test_mobilenetv1/v2 for opencl ci. test=develop