I added a new back end to TVM and have generated the code I need, the code itself can run on the new device, but how should I run it in TVM? If you don’t know, but know how to run cuda source codes generated by my own codegen tool, it might help me too.
