Will I suffer performance drop if I use compiled model on other platforms

1

HI,

I can use tvm to compile my model locally with gpu of Tesla T4, if I want to run inference on another machine with gpu of v100 using the model compiled on T4 gpu, will I experience a performance drop ?