`aten.sym_storage_offset`

[chohk88]

aten op converter for sym_storage_offset

https://github.com/pytorch/pytorch/blob/d0211e207c78fafac2edaf2e14954f668e898b4a/aten/src/ATen/TensorGeometry.h#L88 https://github.com/pytorch/pytorch/blob/d0211e207c78fafac2edaf2e14954f668e898b4a/aten/src/ATen/TensorGeometry.h#L139

[chohk88]

The aten.sym_storage_offset operation returns the offset between the storage start point and the first element of the tensor.

In PyTorch, tensors are allocated their own memory block at creation, naturally starting at 0, and here is an example where the result is not 0.

Therefore, in TensorRT, accessing the memory layout of tensors is impossible, making it infeasible to implement this operation. @narendasan, @gs-olive How should we handle this issue?