* [LITE][OPENCL] improve perf and support variable length for fc_buffer、elementwise_add、scale、activation、grid_sampler. test=develop

* [LITE][OPENCL][Image] add lws turn &  close cl check when shutdownlog , test=develop

* [LITE][OPENCL][Image] add lws turn &  close cl check when shutdownlog , test=develop

* [LITE][OPENCL][Image] add lws turn &  close cl check when shutdownlog , test=develop

* [LITE][OPENCL][Image] add lws turn &  close cl check when shutdownlog , test=develop

* [LITE][OPENCL][Image] add lws turn &  close cl check when shutdownlog , test=develop

* [LITE][BM] add fill_const assign_value ops, test=develop

add cuda kernel.

abs, tanh, elementwise_sub

 fix sequence_pool bug

* [LITE][OPENCL] fix elemul kernel of opencl. test=develop

* fix instanceNorm of opencl. test=develop

* add more info about cl device info. test=develop

* [LITE][OPENCL][Image] mv kernel init gws lws into prepare for run, test=develop

* [LITE][OPENCL][Image] shut down profile, test=develop

* [LITE][OPENCL][Image] move log ahead, 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

* [LITE][OPENCL][Image]optimise conv2d 5x5 7x7,test=develop

* [LITE][OPENCL][Image]optimise conv2d 5x5 7x7,test=develop

* [LITE][OPENCL][Image]optimise conv2d 5x5 7x7,test=develop

* [LITE][OPENCL] comment vlog for opencl kernel. test=develop

* [LITE][OPENCL] clean code for opencl. test=develop

* [LITE][PROFILER] Enhance Precision Profiler. test=develop

* delete useless var in profiler. test=develop

* add ocl header. test=develop

* rm grad code

* add mul_grad, test=develop

* Fix OpenCL kernel of exp, tanh,

* Fix layout pass for opencl backend,

* Add how to debug cl kernel. test=develop.

add cuda c++ doc

- 增加cuda c++ demo.
- 考虑到检测模型尾部一般是multiclass_nms，该kernel为host，如果fetch kernel为cuda的话，则会在此处插入无用的io_copy(host->cuda)，由于该原因，注释掉fetch的cuda kernel. 默认使用host的fetch kernel. 此处暗中进行的行为：每次predictor run完，都会默认把数据从cuda拷贝到cpu

* * support download bm_sdk, test=develop

* [LITE][BM] add slice op

* [LITE][BM] fix concat issue

* [LITE][BM] support reduce full ops, test=develop

* [LITE][BM] change test_resnet50 to change test_classify

* [LITE][BM] add cast op

* [LITE][BM] add reduce and other ops, test=develop

* [LITE][BM] add reduce,cast and other ops, test=develop

add cuda ci.

Organize cuda compilation products.

* add sgd op, test=develop

* test=develop

* add backend

* add kernel

* add grad kernel

* [NPU] subgraph's precision register to kAny

* [LITE][OPENCL] Change fp32 fc to fp16's. test=develop

* fix act in conv3x3opt opencl kernel. test=develop

* [LITE][OPENCL] fix opencl fc kernel. test=develop

* [LITE][OPENCL] add lws for opencl conv image kernel. test=develop