[onert-micro] Add Div kernel

onert-micro/onert-micro/include/pal/common/PALDivCommon.h

@@ -0,0 +1,53 @@
+/*
+ * Copyright (c) 2021 Samsung Electronics Co., Ltd. All Rights Reserved
+ * Copyright 2019 The TensorFlow Authors. All Rights Reserved.
+ *
+ * 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.
+ */
+
+#ifndef ONERT_MICRO_EXECUTE_PAL_DIV_COMMON_H
+#define ONERT_MICRO_EXECUTE_PAL_DIV_COMMON_H
+
+#include "PALArithmeticOpCommon.h"
+
+namespace onert_micro
+{
+namespace execute
+{
+namespace pal
+{
+
+template <typename T>
+OMStatus Div(const core::BinaryArithmeticBroadcastParams &params, const int flat_size,
+             const T *input1_data, const T *input2_data, T *output_data)
+{
+  ArithmeticOp<T, DivFn<T>>(params, flat_size, input1_data, input2_data, output_data);
+  return Ok;
+}
+
+template <typename T>
+OMStatus BroadcastDiv4DSlow(const core::BinaryArithmeticBroadcastParams &params,
+                            const core::OMRuntimeShape &input1_shape, const T *input1_data,
+                            const core::OMRuntimeShape &input2_shape, const T *input2_data,
+                            const core::OMRuntimeShape &output_shape, T *output_data)
+{
+  BroadcastArithmeticOp4DSlow<T, DivFn<T>>(params, input1_shape, input1_data, input2_shape,
+                                           input2_data, output_shape, output_data);
+  return Ok;
+}
+
+} // namespace pal
+} // namespace execute
+} // namespace onert_micro
+
+#endif // ONERT_MICRO_EXECUTE_PAL_MUL_COMMON_H

onert-micro/onert-micro/include/pal/mcu/KernelsToBuild.lst

@@ -10,7 +10,7 @@ REGISTER_KERNEL(BATCH_TO_SPACE_ND, BatchToSpaceND)
 #/*REGISTER_KERNEL(CEIL, Ceil)*/
 REGISTER_KERNEL(COS, Cos)
 #/*REGISTER_KERNEL(CAST, Cast)*/
-#/*REGISTER_KERNEL(DIV, Div)*/
+REGISTER_KERNEL(DIV, Div)
 REGISTER_KERNEL(DEPTHWISE_CONV_2D, DepthwiseConv2D)
 #/*REGISTER_KERNEL(DEPTH_TO_SPACE, DepthToSpace)*/
 #/*REGISTER_KERNEL(DEQUANTIZE, Dequantize)*/

onert-micro/onert-micro/include/pal/mcu/PALDiv.h

@@ -0,0 +1,23 @@
+/*
+ * Copyright (c) 2021 Samsung Electronics Co., Ltd. All Rights Reserved
+ * Copyright 2019 The TensorFlow Authors. All Rights Reserved.
+ *
+ * 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.
+ */
+
+#ifndef ONERT_MICRO_EXECUTE_PAL_DIV_H
+#define ONERT_MICRO_EXECUTE_PAL_DIV_H
+
+#include "PALDivCommon.h"
+
+#endif // ONERT_MICRO_EXECUTE_PAL_MUL_H

onert-micro/onert-micro/include/test_models/div/FloatDivKernel.h

@@ -0,0 +1,184 @@
+/*
+ * Copyright (c) 2024 Samsung Electronics Co., Ltd. All Rights Reserved
+ *
+ * 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.
+ */
+
+#ifndef ONERT_MICRO_TEST_MODELS_DIV_KERNEL_FLOAT_H
+#define ONERT_MICRO_TEST_MODELS_DIV_KERNEL_FLOAT_H
+
+#include "TestDataDivBase.h"
+
+namespace onert_micro
+{
+namespace test_model
+{
+namespace div_float_with_broadcasting
+{
+
+/*
+ * div Kernel:
+ *
+ * Input_1(1, 4, 4, 3)   Input_2(1, 4, 4, 1)
+ *       \             /
+ *     Div(with broadcast)
+ *              |
+ *          Output(1, 4, 4, 3)
+ */
+const unsigned char test_kernel_model_circle[] = {
+  0x18, 0x00, 0x00, 0x00, 0x43, 0x49, 0x52, 0x30, 0x00, 0x00, 0x0e, 0x00, 0x14, 0x00, 0x00, 0x00,
+  0x0c, 0x00, 0x08, 0x00, 0x10, 0x00, 0x04, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00,
+  0x30, 0x00, 0x00, 0x00, 0x6c, 0x01, 0x00, 0x00, 0x88, 0x01, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x1c, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x88, 0xff, 0xff, 0xff, 0x8c, 0xff, 0xff, 0xff, 0x90, 0xff, 0xff, 0xff, 0x94, 0xff, 0xff, 0xff,
+  0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x18, 0x00, 0x14, 0x00,
+  0x10, 0x00, 0x0c, 0x00, 0x08, 0x00, 0x04, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00,
+  0x1c, 0x00, 0x00, 0x00, 0x60, 0x00, 0x00, 0x00, 0x64, 0x00, 0x00, 0x00, 0x6c, 0x00, 0x00, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00,
+  0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x14, 0x00, 0x00, 0x00, 0x10, 0x00, 0x0c, 0x00,
+  0x07, 0x00, 0x08, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1d, 0x10, 0x00, 0x00, 0x00,
+  0x10, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x04, 0x00, 0x04, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00,
+  0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x74, 0x00, 0x00, 0x00,
+  0x34, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0xa4, 0xff, 0xff, 0xff, 0x0c, 0x00, 0x00, 0x00,
+  0x03, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x6f, 0x66, 0x6d, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x03, 0x00, 0x00, 0x00, 0xd0, 0xff, 0xff, 0xff, 0x0c, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00,
+  0x10, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x69, 0x66, 0x6d, 0x32, 0x00, 0x00, 0x00, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x10, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x08, 0x00, 0x04, 0x00,
+  0x0c, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x69, 0x66, 0x6d, 0x31, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x0c, 0x00, 0x0b, 0x00, 0x00, 0x00,
+  0x00, 0x00, 0x04, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2a,
+  0x11, 0x00, 0x00, 0x00, 0x4f, 0x4e, 0x45, 0x2d, 0x74, 0x66, 0x6c, 0x69, 0x74, 0x65, 0x32, 0x63,
+  0x69, 0x72, 0x63, 0x6c, 0x65, 0x00, 0x00, 0x00};
+
+const std::vector<float> input1_data = {
+  0.53560877, -1.8867658, -8.832248,  5.859678,   -5.1357374, -7.014421,  2.8209867,  -4.6921616,
+  8.162995,   -2.6120718, -0.8869447, -2.3489552, -6.208182,  -2.2264383, 6.55287,    -5.3615203,
+  -2.7882025, 1.1029396,  -9.950372,  0.1670622,  3.533298,   2.606648,   7.126593,   -3.6454842,
+  2.637641,   -6.7670174, 5.8970585,  8.792766,   -6.63798,   -1.1457828, 9.122211,   -3.2186837,
+  -8.038172,  3.3369763,  0.89376354, -7.979661,  -4.241393,  -3.4046402, 6.1339216,  7.767886,
+  5.7856345,  8.147999,   9.79144,    8.312661,   -4.9722652, 3.0038614,  -5.3218126, -6.5680394};
+const std::vector<float> input2_data = {
+  -8.79469,  -0.2644837, -1.6307421, 2.0497515, -1.5460981, 1.3044068, 7.1843495, 1.1402534,
+  0.2102437, 4.253287,   7.076496,   1.3749583, 6.6056423,  4.4962177, 3.9589605, -1.8852876};
+const std::vector<float> reference_output_data = {
+  -0.06090138, 0.21453466,  1.0042704,   -22.155157,  19.417974,  26.521185,   -1.7298791,
+  2.8773167,   -5.005694,   -1.2743359,  -0.4327084,  -1.1459707, 4.0153866,   1.4400369,
+  -4.2383275,  -4.110313,   -2.1375253,  0.8455488,   -1.3850067, 0.023253629, 0.49180487,
+  2.286025,    6.250008,    -3.1970825,  12.545636,   -32.18654,  28.048681,   2.0672872,
+  -1.5606706,  -0.26938763, 1.2890859,   -0.45484143, -1.1358972, 2.4269655,   0.6500296,
+  -5.803566,   -0.6420864,  -0.51541394, 0.9285882,   1.727649,   1.2867781,   1.8121896,
+  2.4732351,   2.099708,    -1.2559522,  -1.5933173,  2.822812,   3.4838395};
+
+} // namespace div_float_with_broadcasting
+
+namespace div_float_no_broadcasting
+{
+/*
+ * div Kernel:
+ *
+ * Input_1(1, 4, 4, 3)   Input_2(1, 4, 4, 3)
+ *       \             /
+ *     Div(no broadcast)
+ *              |
+ *          Output(1, 4, 4, 3)
+ */
+const unsigned char test_kernel_model_circle[] = {
+  0x18, 0x00, 0x00, 0x00, 0x43, 0x49, 0x52, 0x30, 0x00, 0x00, 0x0e, 0x00, 0x14, 0x00, 0x00, 0x00,
+  0x0c, 0x00, 0x08, 0x00, 0x10, 0x00, 0x04, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00,
+  0x30, 0x00, 0x00, 0x00, 0x6c, 0x01, 0x00, 0x00, 0x88, 0x01, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x1c, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x88, 0xff, 0xff, 0xff, 0x8c, 0xff, 0xff, 0xff, 0x90, 0xff, 0xff, 0xff, 0x94, 0xff, 0xff, 0xff,
+  0x01, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x18, 0x00, 0x14, 0x00,
+  0x10, 0x00, 0x0c, 0x00, 0x08, 0x00, 0x04, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00,
+  0x1c, 0x00, 0x00, 0x00, 0x60, 0x00, 0x00, 0x00, 0x64, 0x00, 0x00, 0x00, 0x6c, 0x00, 0x00, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x6d, 0x61, 0x69, 0x6e, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00,
+  0x14, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0e, 0x00, 0x14, 0x00, 0x00, 0x00, 0x10, 0x00, 0x0c, 0x00,
+  0x07, 0x00, 0x08, 0x00, 0x0e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1d, 0x10, 0x00, 0x00, 0x00,
+  0x10, 0x00, 0x00, 0x00, 0x14, 0x00, 0x00, 0x00, 0x04, 0x00, 0x04, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00,
+  0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x74, 0x00, 0x00, 0x00,
+  0x34, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0xa4, 0xff, 0xff, 0xff, 0x0c, 0x00, 0x00, 0x00,
+  0x03, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00, 0x6f, 0x66, 0x6d, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x03, 0x00, 0x00, 0x00, 0xd0, 0xff, 0xff, 0xff, 0x0c, 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, 0x00,
+  0x10, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x69, 0x66, 0x6d, 0x32, 0x00, 0x00, 0x00, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x03, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x10, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x08, 0x00, 0x04, 0x00,
+  0x0c, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00,
+  0x04, 0x00, 0x00, 0x00, 0x69, 0x66, 0x6d, 0x31, 0x00, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x03, 0x00, 0x00, 0x00,
+  0x01, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x0c, 0x00, 0x0c, 0x00, 0x0b, 0x00, 0x00, 0x00,
+  0x00, 0x00, 0x04, 0x00, 0x0c, 0x00, 0x00, 0x00, 0x2a, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x2a,
+  0x11, 0x00, 0x00, 0x00, 0x4f, 0x4e, 0x45, 0x2d, 0x74, 0x66, 0x6c, 0x69, 0x74, 0x65, 0x32, 0x63,
+  0x69, 0x72, 0x63, 0x6c, 0x65, 0x00, 0x00, 0x00};
+
+std::vector<float> input1_data = {
+  -0.45486993, 7.7920837,  2.458276,   -9.338201, 2.132607,   7.5134172,   -7.3420544, -6.390925,
+  1.456884,    -1.8421048, 2.4459255,  -3.930421, 5.439674,   -0.3640652,  -4.5039334, -8.457781,
+  3.8367634,   -6.271797,  -3.6314914, 7.7209616, 0.7935352,  -6.561332,   0.77153474, 8.639922,
+  0.15612718,  5.9307566,  0.71869296, 5.4473047, -7.4552007, -2.4453044,  -5.7700257, 7.9675865,
+  4.7339244,   1.3709338,  6.940702,   2.7984607, 9.185996,   8.848705,    -4.7602606, -3.887096,
+  4.807505,    2.9972825,  5.416388,   5.6682634, -5.875255,  -0.29188222, 7.522483,   -6.873605};
+std::vector<float> input2_data = {
+  1.5015899,  9.9904785,  -5.368375,  -3.0589666, 1.4541256,  3.0067844,  -1.0564902,  3.4706006,
+  4.3666296,  9.586457,   -4.497033,  -9.208905,  -6.8920064, 6.2652717,  -0.60215276, 7.8108196,
+  -7.7110577, -4.182856,  6.8893385,  9.817479,   -9.451469,  -2.0223913, 5.1733356,   5.4378557,
+  4.1402535,  -4.4087567, 0.16630036, -8.007953,  4.7895074,  -3.2770896, 4.013145,    2.2894316,
+  5.74741,    -9.481342,  3.194859,   9.471764,   -7.8339133, -8.103412,  -3.8862576,  9.901491,
+  5.358148,   -2.7227716, -5.84558,   9.881574,   3.5261114,  3.4573863,  8.221204,    -3.6716719};
+std::vector<float> reference_output_data = {
+  -0.30292553, 0.779951,   -0.4579181,  3.0527303,   1.4665906,   2.4988215,   6.949477,
+  -1.8414464,  0.3336404,  -0.19215699, -0.54389757, 0.42680657,  -0.7892729,  -0.058108445,
+  7.4797187,   -1.0828289, -0.49756643, 1.4994054,   -0.52711755, 0.7864505,   -0.083958924,
+  3.2443435,   0.14913681, 1.5888473,   0.03770957,  -1.345222,   4.321656,    -0.6802369,
+  -1.5565695,  0.7461817,  -1.4377816,  3.4801593,   0.8236622,   -0.14459279, 2.1724596,
+  0.29545298,  -1.1725935, -1.0919728,  1.2248958,   -0.3925768,  0.8972326,   -1.1008204,
+  -0.92657834, 0.5736195,  -1.6662138,  -0.0844228,  0.9150099,   1.872064};
+
+} // namespace div_float_no_broadcasting
+
+class TestDataFloatDiv : public TestDataDivBase<float>
+{
+public:
+  explicit TestDataFloatDiv(bool is_with_broadcast) : TestDataDivBase<float>(is_with_broadcast)
+  {
+    if (is_with_broadcast)
+    {
+      _input1_data = div_float_with_broadcasting::input1_data;
+      _input2_data = div_float_with_broadcasting::input2_data;
+      _reference_output_data = div_float_with_broadcasting::reference_output_data;
+      _test_kernel_model_circle = div_float_with_broadcasting::test_kernel_model_circle;
+    }
+    else
+    {
+      _input1_data = div_float_no_broadcasting::input1_data;
+      _input2_data = div_float_no_broadcasting::input2_data;
+      _reference_output_data = div_float_no_broadcasting::reference_output_data;
+      _test_kernel_model_circle = div_float_no_broadcasting::test_kernel_model_circle;
+    }
+  }
+
+  ~TestDataFloatDiv() override = default;
+};
+
+} // namespace test_model
+} // namespace onert_micro
+
+#endif // ONERT_MICRO_TEST_MODELS_MUL_KERNEL_FLOAT_H