use float32 for mad to improve precision

tests/common/test_op/matmul.py

@@ -21,6 +21,7 @@ from akg import backend as cce
 from akg.utils import kernel_exec as utils
 from akg.utils import custom_tiling as ct_util
 from akg.utils import validation_check as vc_util
+from akg.ops.math import cast
 
 logging.basicConfig(level=logging.DEBUG)
 
@@ -166,7 +167,7 @@ def matmul4D_compute(x, y, bias_value, out_dtype, left_format, right_format, out
                 if adj_y:
                     y_indices = indices[:(N - 4)] + (ko,) + indices[(N - 4):(N - 3)] + indices[(N - 1):] + (ki,)
 
-        return akg.lang.cce.mmad((x(*x_indices) * y(*y_indices)).astype(out_dtype), axis=[ko, ki])
+        return akg.lang.cce.mmad((x(*x_indices) * y(*y_indices)).astype("float32"), axis=[ko, ki])
 
 
     if left_format == "zZ":
@@ -223,6 +224,8 @@ def matmul4D_compute(x, y, bias_value, out_dtype, left_format, right_format, out
         "bias": bias_name,
     })
 
+    if out_dtype == "float16":
+        result_matmul = cast.cast(result_matmul, out_dtype)
 
     def matmul_reshape(shape, result_matmul, *indices):
         N = len(shape)

tests/operators/cube/test_matmul4d_ad_001.py

-# Copyright 2019 Huawei Technologies Co., Ltd
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-# http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-"""
-matmul4d_ad
-"""
-import datetime
-import os
-import pytest
-from base import TestBase
-from nose.plugins.attrib import attr
-from test_run.matmul4d_ad_run import matmul4d_ad_run
-
-
-class TestCase(TestBase):
-
-    def setup(self):
-        case_name = "test_akg_matmul_001"
-        case_path = os.getcwd()
-        self.params_init(case_name, case_path)
-        self.caseresult = True
-        self._log.info("============= {0} Setup case============".format(self.casename))
-        self.testarg = [
-            # caseflag, opfuncname, testRunArgs, dimArgs
-            # shape_x, shape_y, bias, bypass, adj_x, adj_y, dtype, out_dtype, kernel_name, attrs
-            ("matmul4d_ad_run_0", matmul4d_ad_run, ((64, 128), (128, 32), 0,   False, False,
-                                                    "float16", "float16", "matmul4d_ad_cce")),
-            ("matmul4d_ad_run_1", matmul4d_ad_run, ((64, 1024), (1024, 32), 0,   False, False,
-                                                    "float16", "float16", "matmul4d_ad_cce")),
-            ("matmul4d_ad_run_2", matmul4d_ad_run, ((1024, 64), (1024, 32), 0,  True, False,
-                                                    "float16", "float16", "matmul4d_ad_cce")),
-            ("matmul4d_ad_run_3", matmul4d_ad_run, ((64, 1024), (32, 1024), 0,   False, True,
-                                                    "float16", "float16", "matmul4d_ad_cce")),
-            ("matmul4d_ad_run_4", matmul4d_ad_run, ((1024, 64), (32, 1024), 0, True, True,
-                                                    "float16", "float16", "matmul4d_ad_cce")),
-        ]
-        return
-
-    @pytest.mark.rpc_mini
-    @pytest.mark.level0
-    @pytest.mark.env_onecard
-    @pytest.mark.platform_x86_ascend_training
-    def test_run(self):
-        """
-        run case.#
-        :return:
-        """
-        self.common_run(self.testarg)
-
-    def teardown(self):
-        """
-        clean environment
-        :return:
-        """
-        self._log.info("============= {0} Teardown============".format(self.casename))
-        return
-
-if __name__ == "__main__":
-    a = TestCase()
-    a.setup()
-    a.test_run()