How to let tvm_runtime deploy in c++(gcc-4.8)

#1

I set build_static runtime “ON” but some errors still happend.I built tvm with llvm and cuda.