Egaerly execute inplace ops if in eager mode

[JackCaoG]

Even with functionization we still see some inplace ops, like optimization_barrier_, all_reduce_, currently in eager mode they won't be execute.

The eager mode today works by

xla/torch_xla/csrc/tensor.cpp

Lines 81 to 93 in 98dd99e

	XLATensorPtr XLATensor::Create(
	torch::lazy::Value ir_value, const torch::lazy::BackendDevice& device,
	std::optional<at::ScalarType> logical_element_type) {
	XLATensorPtr xtensor = c10::make_intrusive<XLATensor>(
	XLATensor(std::move(ir_value), device, logical_element_type));
	XLAGraphExecutor* graph_executor = XLAGraphExecutor::Get();
	graph_executor->RegisterTensor(xtensor->data());
	if (UseEagerDebugMode() || graph_executor->UseEagerMode()) {
	std::vector<XLATensorPtr> xtensors({xtensor});
	graph_executor->ApplyEagerSync(xtensors);
	}
	return xtensor;
	}

when creating a new XLATensor with a IR, we will execute that IR. This doesn't handle the inplace update cases since there is no new XLATensor being created.

I also need to handle the two special cases in follow up pr

1. random seed IR
2. all_reduce_token IR