TVM deep learning compiler stack

Did a brief performance study on TVM deep learning compiler stack for different hardware backends ranging from x86 based CPUs, ARM based CPUs and integrated GPUs and Xilinx FPGAs