Merge develop to release1.2 branch

Fix ios cross compile, revert quantize kernel, and add other dequantize fusion pattern

Use user mode driver for FPGA track close #1325

imp googlenet for gpu

fix a bug in format_conv_data function

Fix quantize kernel while pad != 0, implement the fusion kernel of dequantize, elementwise_add, batch norm and Relu

fix a bug in feed kernel

Optimize im2col and implement winograd conv, bring 35% speed up for googlenet

add tanh kernel in FPGA v1 fixed #1314

fix #1310  update android development document

add split and transpose kernels for FPGA track close #1312