Rename use_mkldnn to use_onednn in test cases - part2 (#73902)

Author: co63oc

test/mkldnn/test_sum_bf16_mkldnn_op.py

@@ -28,8 +28,8 @@
 class TestSumBF16ONEDNN(TestSumOp):
  def setUp(self):
  self.op_type = "sum"
- self.use_mkldnn = True
- self.mkldnn_data_type = "bfloat16"
+ self.use_onednn = True
+ self.onednn_data_type = "bfloat16"
 
  # float32 input to be use for reference
  x0 = np.random.random((25, 8)).astype('float32')
@@ -45,7 +45,7 @@ def setUp(self):
 
  y = x0 + x1 + x2
  self.outputs = {'Out': convert_float_to_uint16(y)}
- self.attrs = {'use_mkldnn': self.use_mkldnn}
+ self.attrs = {'use_mkldnn': self.use_onednn}
 
  def test_check_output(self):
  self.check_output_with_place(core.CPUPlace(), check_pir_onednn=True)