[boo driver] Multidevice benchmarking doesn't work

[rkayaith]

--gpu-id -1 is broken:

~/repos/iree-turbine$ ROCR_VISIBLE_DEVICES=0,1 iree-boo-driver conv --gpu-id -1

>>> conv

Traceback (most recent call last):
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/kernel/boo/driver/driver.py", line 188, in main
    prof = run(
           ^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/kernel/boo/driver/driver.py", line 285, in run
    example_results = func(*per_device_args[0])
                      ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/kernel/boo/exports/signature.py", line 76, in run_and_verify
    result = compiled(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_dynamo/eval_frame.py", line 414, in __call__
    return super().__call__(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_dynamo/eval_frame.py", line 832, in compile_wrapper
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/kernel/boo/op_exports/conv.py", line 442, in forward
    def forward(self, *args: torch.Tensor) -> torch.Tensor:
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_dynamo/eval_frame.py", line 1044, in _fn
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_functorch/aot_autograd.py", line 1130, in forward
    return compiled_fn(full_args)
           ^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_functorch/_aot_autograd/runtime_wrappers.py", line 353, in runtime_wrapper
    all_outs = call_func_at_runtime_with_args(
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_functorch/_aot_autograd/utils.py", line 129, in call_func_at_runtime_with_args
    out = normalize_as_list(f(args))
                            ^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_functorch/_aot_autograd/runtime_wrappers.py", line 724, in inner_fn
    outs = compiled_fn(args)
           ^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_functorch/_aot_autograd/runtime_wrappers.py", line 526, in wrapper
    return compiled_fn(runtime_args)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_functorch/_aot_autograd/utils.py", line 103, in g
    return f(*args)
           ^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/fx/_lazy_graph_module.py", line 126, in _lazy_forward
    return self(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/fx/graph_module.py", line 837, in call_wrapped
    return self._wrapped_call(self, *args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/fx/graph_module.py", line 413, in __call__
    raise e
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/fx/graph_module.py", line 400, in __call__
    return super(self.cls, obj).__call__(*args, **kwargs)  # type: ignore[misc]
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/nn/modules/module.py", line 1786, in _call_impl
    return forward_call(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "<eval_with_key>.5 from /home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/fx/experimental/proxy_tensor.py:1350 in wrapped", line 7, in forward
    fused_op_layout_cus_b2e5e1ff62d0dafe94e32d1e47e3fd273cc15dd5 = torch.ops.boo.fused_op_layout_cus_b2e5e1ff62d0dafe94e32d1e47e3fd273cc15dd5(permute, permute_1);  permute = permute_1 = None
                                                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_ops.py", line 1255, in __call__
    return self._op(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/kernel/boo/ops/graph.py", line 314, in _
    outputs = launch(*[arg.data for arg in handled_inputs])
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/runtime/launch.py", line 230, in __call__
    arg_list.push_ref(turbine_device.import_torch_tensor(arg))
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/runtime/device.py", line 405, in <lambda>
    self.import_torch_tensor = lambda t: import_fn(self, t)
                                         ^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/iree-turbine/iree/turbine/runtime/device.py", line 551, in _device_import_torch_tensor_cuda_hip
    capsule = t.__dlpack__(stream=None)
              ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/rkayaith/repos/.venv/lib/python3.11/site-packages/torch/_tensor.py", line 1726, in __dlpack__
    raise BufferError(
BufferError: Can't export tensors on a different CUDA device index. Expected: 0. Current device: 1.

I believe this is related to the bump to pytorch 2.9. There may be additional issues with the iree_boo_experimental backend, as it hasn't been testsed.