[Break XPU][Inductor UT] Generalize device-bias code introduced by pytorch#146866. (pytorch#148534)

Pull Request resolved: pytorch#148534
Approved by: https://github.com/nandesuka

Author: etaf

test/inductor/test_padding.py

@@ -760,7 +760,7 @@ def get_input(size: tuple[int], alignment_bytes: int) -> torch.Tensor:
 
         output_shape = (shape[0] * num_inputs, shape[1])
         output_stride = input_tensors[0].stride()
-        output_line = f"buf12 = empty_strided_cuda({output_shape}, {output_stride}, torch.float32)"
+        output_line = f"buf12 = empty_strided_{GPU_TYPE}({output_shape}, {output_stride}, torch.float32)"
         self.assertTrue(any(output_line in line for line in code))