TVM's runtime is designed to be lightweight and portable. There are several ways you can integrate TVM into your project.
This article introduces possible ways to integrate TVM as a JIT compiler to generate functions on your system.
TVM's generated function follows the PackedFunc convention. It is a function that can take positional arguments including standard types such as float, integer, string. The PackedFunc takes DLTensor pointer in dlpack convention. So the only thing you need to solve is to create a corresponding DLTensor object.
The only thing we have to do in C++ is to convert your array to DLTensor and pass in its address as DLTensor*
to the generated function.
Assume you have a python object MyArray
. There are three things that you need to do
_tvm_tcode
field to your array which returns tvm.TypeCode.ARRAY_HANDLE
_tvm_handle
property in your object, which returns the address of DLTensor in python integertvm.register_extension
# Example code import tvm class MyArray(object): _tvm_tcode = tvm.TypeCode.ARRAY_HANDLE @property def _tvm_handle(self): dltensor_addr = self.get_dltensor_addr() return dltensor_addr # You can put registration step in a separate file mypkg.tvm.py # and only optionally import that if you only want optional dependency. tvm.register_extension(MyArray)