Star1289
1398
代码
代码
Issues95
Pull Requests183
Actions
项目讨论
Wiki
分析
Star1289
1398
已合并
950支持静态库功能检查与修正 #1872
Sun创建于 3月25日
950支持静态库功能检查与修正 #1872
已合并
Sun创建于 3月25日
从LuckySun/ops-mathstatic合入到cann/ops-mathmaster
共 50 个文件变更+274-280
Mcommon/inc/op_host/tiling_templates_registry.h+0-6
@@ -325,12 +325,6 @@ private:
325 static Ops::Math::OpTiling::RegisterNew VAR_UNUSED##op_type##class_name##priority_register = \325 static Ops::Math::OpTiling::RegisterNew VAR_UNUSED##op_type##class_name##priority_register = \
326 Ops::Math::OpTiling::RegisterNew(#op_type).tiling<class_name>(priority, soc_versions)326 Ops::Math::OpTiling::RegisterNew(#op_type).tiling<class_name>(priority, soc_versions)
327 327 
328-// op_type: 算子名称， class_name: 注册的 tiling 类,
329-// priority: tiling 类的优先级, 越小表示优先级越高, 即被选中的概率越大
330-#define REGISTER_TILING_TEMPLATE(op_type, class_name, priority) \
331- static Ops::Math::OpTiling::Register VAR_UNUSED##op_type_##class_name##priority_register = \
332- Ops::Math::OpTiling::Register(op_type).tiling<class_name>(priority)
333- 
334// op_type: 算子名称， class_name: 注册的 tiling 类,328// op_type: 算子名称， class_name: 注册的 tiling 类,
335// soc_version: soc版本，用于区分不同的soc329// soc_version: soc版本，用于区分不同的soc
336// priority: tiling 类的优先级, 越小表示优先级越高, 即会优先选择这个tiling类330// priority: tiling 类的优先级, 越小表示优先级越高, 即会优先选择这个tiling类
Mconversion/clip_by_value_v2/examples/test_aclnn_clamp_max.cpp+6-6
@@ -76,17 +76,17 @@ int PrepareInputAndOutput(
76 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self,76 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self,
77 aclScalar** max, void** outDeviceAddr, aclTensor** out)77 aclScalar** max, void** outDeviceAddr, aclTensor** out)
78{78{
79- std::vector<int8_t> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};79+ std::vector<int32_t> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
80- std::vector<int8_t> outHostData(8, 0);80+ std::vector<int32_t> outHostData(8, 0);
81- int8_t maxValue = 4;81+ int32_t maxValue = 4;
82 // 创建self aclTensor82 // 创建self aclTensor
83- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_INT8, self);83+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_INT32, self);
84 CHECK_RET(ret == ACL_SUCCESS, return ret);84 CHECK_RET(ret == ACL_SUCCESS, return ret);
85 // 创建max aclScalar85 // 创建max aclScalar
86- *max = aclCreateScalar(&maxValue, aclDataType::ACL_INT8);86+ *max = aclCreateScalar(&maxValue, aclDataType::ACL_INT32);
87 CHECK_RET(*max != nullptr, return ret);87 CHECK_RET(*max != nullptr, return ret);
88 // 创建out aclTensor88 // 创建out aclTensor
89- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_INT8, out);89+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_INT32, out);
90 CHECK_RET(ret == ACL_SUCCESS, return ret);90 CHECK_RET(ret == ACL_SUCCESS, return ret);
91 91 
92 return ACL_SUCCESS;92 return ACL_SUCCESS;
Mconversion/clip_by_value_v2/examples/test_aclnn_clamp_max_tensor.cpp+7-7
@@ -77,18 +77,18 @@ int PrepareInputAndOutput(
77 void** selfDeviceAddr, aclTensor** self, void** maxDeviceAddr, aclTensor** max, void** outDeviceAddr,77 void** selfDeviceAddr, aclTensor** self, void** maxDeviceAddr, aclTensor** max, void** outDeviceAddr,
78 aclTensor** out)78 aclTensor** out)
79{79{
80- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};80+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
81- std::vector<double> maxHostData = {1, 1, 1, 2, 2, 2, 3, 3};81+ std::vector<float> maxHostData = {1, 1, 1, 2, 2, 2, 3, 3};
82- std::vector<double> outHostData(8, 0);82+ std::vector<float> outHostData(8, 0);
83 83 
84 // 创建self aclTensor84 // 创建self aclTensor
85- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);85+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87 // 创建max aclTensor87 // 创建max aclTensor
88- ret = CreateAclTensor(maxHostData, maxShape, maxDeviceAddr, aclDataType::ACL_DOUBLE, max);88+ ret = CreateAclTensor(maxHostData, maxShape, maxDeviceAddr, aclDataType::ACL_FLOAT, max);
89 CHECK_RET(ret == ACL_SUCCESS, return ret);89 CHECK_RET(ret == ACL_SUCCESS, return ret);
90 // 创建out aclTensor90 // 创建out aclTensor
91- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);91+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
92 CHECK_RET(ret == ACL_SUCCESS, return ret);92 CHECK_RET(ret == ACL_SUCCESS, return ret);
93 93 
94 return ACL_SUCCESS;94 return ACL_SUCCESS;
@@ -163,7 +163,7 @@ int main()
163 163 
164 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改164 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
165 auto size = GetShapeSize(outShape);165 auto size = GetShapeSize(outShape);
166- std::vector<double> resultData(size, 0);166+ std::vector<float> resultData(size, 0);
167 ret = aclrtMemcpy(167 ret = aclrtMemcpy(
168 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),168 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),
169 ACL_MEMCPY_DEVICE_TO_HOST);169 ACL_MEMCPY_DEVICE_TO_HOST);
Mconversion/clip_by_value_v2/examples/test_aclnn_clamp_min.cpp+8-8
@@ -77,19 +77,19 @@ int PrepareInputAndOutput(
77 std::vector<int64_t>& shape, void** selfDeviceAddr, aclTensor** self, aclScalar** min, void** outDeviceAddr,77 std::vector<int64_t>& shape, void** selfDeviceAddr, aclTensor** self, aclScalar** min, void** outDeviceAddr,
78 aclTensor** out)78 aclTensor** out)
79{79{
80- double min_v = 2;80+ float min_v = 2;
81 81 
82- std::vector<double> selfHostData = {0, 1, 0, 3, 0, 5, 0, 7};82+ std::vector<float> selfHostData = {0, 1, 0, 3, 0, 5, 0, 7};
83- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};83+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
84 84 
85 // 创建self aclTensor85 // 创建self aclTensor
86- auto ret = CreateAclTensor(selfHostData, shape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);86+ auto ret = CreateAclTensor(selfHostData, shape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
87 CHECK_RET(ret == ACL_SUCCESS, return ret);87 CHECK_RET(ret == ACL_SUCCESS, return ret);
88 // 创建min88 // 创建min
89- *min = aclCreateScalar(&min_v, aclDataType::ACL_DOUBLE);89+ *min = aclCreateScalar(&min_v, aclDataType::ACL_FLOAT);
90 CHECK_RET(*min != nullptr, return ret);90 CHECK_RET(*min != nullptr, return ret);
91 // 创建out aclTensor91 // 创建out aclTensor
92- ret = CreateAclTensor(outHostData, shape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);92+ ret = CreateAclTensor(outHostData, shape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 94 
95 return ACL_SUCCESS;95 return ACL_SUCCESS;
@@ -157,9 +157,9 @@ int main()
157 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);157 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
158 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改158 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
159 auto size = GetShapeSize(shape);159 auto size = GetShapeSize(shape);
160- std::vector<double> resultData(size, 0);160+ std::vector<float> resultData(size, 0);
161 ret = aclrtMemcpy(161 ret = aclrtMemcpy(
162- resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(double),162+ resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(float),
163 ACL_MEMCPY_DEVICE_TO_HOST);163 ACL_MEMCPY_DEVICE_TO_HOST);
164 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);164 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
165 165 
Mconversion/clip_by_value_v2/examples/test_aclnn_clamp_min_tensor.cpp+8-8
@@ -78,17 +78,17 @@ int PrepareInputAndOutput(
78 void** selfDeviceAddr, aclTensor** self, void** minDeviceAddr, aclTensor** min, void** outDeviceAddr,78 void** selfDeviceAddr, aclTensor** self, void** minDeviceAddr, aclTensor** min, void** outDeviceAddr,
79 aclTensor** out)79 aclTensor** out)
80{80{
81- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};81+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
82- std::vector<double> minHostData = {2, 1, 1, 2, 2, 6, 6, 9};82+ std::vector<float> minHostData = {2, 1, 1, 2, 2, 6, 6, 9};
83- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};83+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
84 // 创建self aclTensor84 // 创建self aclTensor
85- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);85+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87 // 创建min aclTensor87 // 创建min aclTensor
88- ret = CreateAclTensor(minHostData, minShape, minDeviceAddr, aclDataType::ACL_DOUBLE, min);88+ ret = CreateAclTensor(minHostData, minShape, minDeviceAddr, aclDataType::ACL_FLOAT, min);
89 CHECK_RET(ret == ACL_SUCCESS, return ret);89 CHECK_RET(ret == ACL_SUCCESS, return ret);
90 // 创建out aclTensor90 // 创建out aclTensor
91- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);91+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
92 CHECK_RET(ret == ACL_SUCCESS, return ret);92 CHECK_RET(ret == ACL_SUCCESS, return ret);
93 93 
94 return ACL_SUCCESS;94 return ACL_SUCCESS;
@@ -161,9 +161,9 @@ int main()
161 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);161 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
162 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改162 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
163 auto size = GetShapeSize(outShape);163 auto size = GetShapeSize(outShape);
164- std::vector<double> resultData(size, 0);164+ std::vector<float> resultData(size, 0);
165 ret = aclrtMemcpy(165 ret = aclrtMemcpy(
166- resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(double),166+ resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(float),
167 ACL_MEMCPY_DEVICE_TO_HOST);167 ACL_MEMCPY_DEVICE_TO_HOST);
168 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);168 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
169 for (int64_t i = 0; i < size; i++) {169 for (int64_t i = 0; i < size; i++) {
Mconversion/clip_by_value_v2/examples/test_aclnn_clamp_tensor.cpp+8-8
@@ -77,22 +77,22 @@ int PrepareInputAndOutput(
77 std::vector<int64_t>& shape, void** selfDeviceAddr, aclTensor** self, void** minDeviceAddr,77 std::vector<int64_t>& shape, void** selfDeviceAddr, aclTensor** self, void** minDeviceAddr,
78 aclTensor** clipValueMin, void** maxDeviceAddr, aclTensor** clipValueMax, void** outDeviceAddr, aclTensor** out)78 aclTensor** clipValueMin, void** maxDeviceAddr, aclTensor** clipValueMax, void** outDeviceAddr, aclTensor** out)
79{79{
80- std::vector<int8_t> selfHostData = {0, 1, 0, 3, 0, 5, 0, 7};80+ std::vector<int32_t> selfHostData = {0, 1, 0, 3, 0, 5, 0, 7};
81- std::vector<int8_t> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};81+ std::vector<int32_t> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
82- std::vector<int8_t> minHostData = {1, 3, 0, 0, 0, 0, 0, 0};82+ std::vector<int32_t> minHostData = {1, 3, 0, 0, 0, 0, 0, 0};
83- std::vector<int8_t> maxHostData = {5, 5, 3, 3, 4, 5, 6, 6};83+ std::vector<int32_t> maxHostData = {5, 5, 3, 3, 4, 5, 6, 6};
84 84 
85 // 创建self aclTensor85 // 创建self aclTensor
86- auto ret = CreateAclTensor(selfHostData, shape, selfDeviceAddr, aclDataType::ACL_INT8, self);86+ auto ret = CreateAclTensor(selfHostData, shape, selfDeviceAddr, aclDataType::ACL_INT32, self);
87 CHECK_RET(ret == ACL_SUCCESS, return ret);87 CHECK_RET(ret == ACL_SUCCESS, return ret);
88 // 创建min aclTensor88 // 创建min aclTensor
89- ret = CreateAclTensor(minHostData, shape, minDeviceAddr, aclDataType::ACL_INT8, clipValueMin);89+ ret = CreateAclTensor(minHostData, shape, minDeviceAddr, aclDataType::ACL_INT32, clipValueMin);
90 CHECK_RET(ret == ACL_SUCCESS, return ret);90 CHECK_RET(ret == ACL_SUCCESS, return ret);
91 // 创建max aclTensor91 // 创建max aclTensor
92- ret = CreateAclTensor(maxHostData, shape, maxDeviceAddr, aclDataType::ACL_INT8, clipValueMax);92+ ret = CreateAclTensor(maxHostData, shape, maxDeviceAddr, aclDataType::ACL_INT32, clipValueMax);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 // 创建out aclTensor94 // 创建out aclTensor
95- ret = CreateAclTensor(outHostData, shape, outDeviceAddr, aclDataType::ACL_INT8, out);95+ ret = CreateAclTensor(outHostData, shape, outDeviceAddr, aclDataType::ACL_INT32, out);
96 CHECK_RET(ret == ACL_SUCCESS, return ret);96 CHECK_RET(ret == ACL_SUCCESS, return ret);
97 97 
98 return ACL_SUCCESS;98 return ACL_SUCCESS;
Mconversion/clip_by_value_v2/examples/test_aclnn_hardtanh.cpp+10-10
@@ -77,21 +77,21 @@ int PrepareInputAndOutput(
77 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self,77 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self,
78 aclScalar** clipValueMin, aclScalar** clipValueMax, void** outDeviceAddr, aclTensor** out)78 aclScalar** clipValueMin, aclScalar** clipValueMax, void** outDeviceAddr, aclTensor** out)
79{79{
80- std::vector<double> selfHostData = {0, 1, 2, 3};80+ std::vector<float> selfHostData = {0, 1, 2, 3};
81- std::vector<double> outHostData = {0, 0, 0, 0};81+ std::vector<float> outHostData = {0, 0, 0, 0};
82- double clipValueMinValue = 1.2;82+ float clipValueMinValue = 1.2;
83- double clipValueMaxValue = 2.4;83+ float clipValueMaxValue = 2.4;
84 // 创建self aclTensor84 // 创建self aclTensor
85- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);85+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87 // 创建clipValueMin aclScalar87 // 创建clipValueMin aclScalar
88- *clipValueMin = aclCreateScalar(&clipValueMinValue, aclDataType::ACL_DOUBLE);88+ *clipValueMin = aclCreateScalar(&clipValueMinValue, aclDataType::ACL_FLOAT);
89 CHECK_RET(*clipValueMin != nullptr, return ret);89 CHECK_RET(*clipValueMin != nullptr, return ret);
90 // 创建clipValueMax aclScalar90 // 创建clipValueMax aclScalar
91- *clipValueMax = aclCreateScalar(&clipValueMaxValue, aclDataType::ACL_DOUBLE);91+ *clipValueMax = aclCreateScalar(&clipValueMaxValue, aclDataType::ACL_FLOAT);
92 CHECK_RET(*clipValueMax != nullptr, return ret);92 CHECK_RET(*clipValueMax != nullptr, return ret);
93 // 创建out aclTensor93 // 创建out aclTensor
94- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);94+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 96 
97 return ACL_SUCCESS;97 return ACL_SUCCESS;
@@ -165,9 +165,9 @@ int main()
165 165 
166 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改166 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
167 auto size = GetShapeSize(outShape);167 auto size = GetShapeSize(outShape);
168- std::vector<double> resultData(size, 0);168+ std::vector<float> resultData(size, 0);
169 ret = aclrtMemcpy(169 ret = aclrtMemcpy(
170- resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(double),170+ resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(float),
171 ACL_MEMCPY_DEVICE_TO_HOST);171 ACL_MEMCPY_DEVICE_TO_HOST);
172 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);172 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
173 for (int64_t i = 0; i < size; i++) {173 for (int64_t i = 0; i < size; i++) {
Mconversion/clip_by_value_v2/examples/test_aclnn_inplace_ardtanh.cpp+10-10
@@ -77,21 +77,21 @@ int PrepareInputAndOutput(
77 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self,77 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self,
78 aclScalar** clipValueMin, aclScalar** clipValueMax, void** outDeviceAddr, aclTensor** out)78 aclScalar** clipValueMin, aclScalar** clipValueMax, void** outDeviceAddr, aclTensor** out)
79{79{
80- std::vector<double> selfHostData = {0, 1, 2, 3};80+ std::vector<float> selfHostData = {0, 1, 2, 3};
81- std::vector<double> outHostData = {0, 0, 0, 0};81+ std::vector<float> outHostData = {0, 0, 0, 0};
82- double clipValueMinValue = 1.2;82+ float clipValueMinValue = 1.2;
83- double clipValueMaxValue = 2.4;83+ float clipValueMaxValue = 2.4;
84 // 创建self aclTensor84 // 创建self aclTensor
85- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);85+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87 // 创建clipValueMin aclScalar87 // 创建clipValueMin aclScalar
88- *clipValueMin = aclCreateScalar(&clipValueMinValue, aclDataType::ACL_DOUBLE);88+ *clipValueMin = aclCreateScalar(&clipValueMinValue, aclDataType::ACL_FLOAT);
89 CHECK_RET(*clipValueMin != nullptr, return ret);89 CHECK_RET(*clipValueMin != nullptr, return ret);
90 // 创建clipValueMax aclScalar90 // 创建clipValueMax aclScalar
91- *clipValueMax = aclCreateScalar(&clipValueMaxValue, aclDataType::ACL_DOUBLE);91+ *clipValueMax = aclCreateScalar(&clipValueMaxValue, aclDataType::ACL_FLOAT);
92 CHECK_RET(*clipValueMax != nullptr, return ret);92 CHECK_RET(*clipValueMax != nullptr, return ret);
93 // 创建out aclTensor93 // 创建out aclTensor
94- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);94+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 96 
97 return ACL_SUCCESS;97 return ACL_SUCCESS;
@@ -168,9 +168,9 @@ int main()
168 168 
169 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改169 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
170 auto size = GetShapeSize(outShape);170 auto size = GetShapeSize(outShape);
171- std::vector<double> resultData(size, 0);171+ std::vector<float> resultData(size, 0);
172 ret = aclrtMemcpy(172 ret = aclrtMemcpy(
173- resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(double),173+ resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(float),
174 ACL_MEMCPY_DEVICE_TO_HOST);174 ACL_MEMCPY_DEVICE_TO_HOST);
175 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);175 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
176 for (int64_t i = 0; i < size; i++) {176 for (int64_t i = 0; i < size; i++) {
Mconversion/clip_by_value_v2/examples/test_aclnn_inplace_clamp_max.cpp+4-4
@@ -74,13 +74,13 @@ int CreateAclTensor(
74 74 
75int PrepareInputAndOutput(std::vector<int64_t>& selfShape, void** selfDeviceAddr, aclTensor** self, aclScalar** max)75int PrepareInputAndOutput(std::vector<int64_t>& selfShape, void** selfDeviceAddr, aclTensor** self, aclScalar** max)
76{76{
77- std::vector<int8_t> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};77+ std::vector<int32_t> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
78- int8_t maxValue = 4;78+ int32_t maxValue = 4;
79 // 创建self aclTensor79 // 创建self aclTensor
80- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_INT8, self);80+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_INT32, self);
81 CHECK_RET(ret == ACL_SUCCESS, return ret);81 CHECK_RET(ret == ACL_SUCCESS, return ret);
82 // 创建max aclScalar82 // 创建max aclScalar
83- *max = aclCreateScalar(&maxValue, aclDataType::ACL_INT8);83+ *max = aclCreateScalar(&maxValue, aclDataType::ACL_INT32);
84 CHECK_RET(*max != nullptr, return ret);84 CHECK_RET(*max != nullptr, return ret);
85 85 
86 return ACL_SUCCESS;86 return ACL_SUCCESS;
Mconversion/clip_by_value_v2/examples/test_aclnn_inplace_clamp_max_tensor.cpp+5-5
@@ -76,13 +76,13 @@ int PrepareInputAndOutput(
76 std::vector<int64_t>& selfShape, std::vector<int64_t>& maxShape, void** selfDeviceAddr, aclTensor** self,76 std::vector<int64_t>& selfShape, std::vector<int64_t>& maxShape, void** selfDeviceAddr, aclTensor** self,
77 void** maxDeviceAddr, aclTensor** max)77 void** maxDeviceAddr, aclTensor** max)
78{78{
79- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};79+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
80- std::vector<double> maxHostData = {1, 1, 1, 2, 2, 2, 3, 3};80+ std::vector<float> maxHostData = {1, 1, 1, 2, 2, 2, 3, 3};
81 // 创建self aclTensor81 // 创建self aclTensor
82- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);82+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
83 CHECK_RET(ret == ACL_SUCCESS, return ret);83 CHECK_RET(ret == ACL_SUCCESS, return ret);
84 // 创建max aclTensor84 // 创建max aclTensor
85- ret = CreateAclTensor(maxHostData, maxShape, maxDeviceAddr, aclDataType::ACL_DOUBLE, max);85+ ret = CreateAclTensor(maxHostData, maxShape, maxDeviceAddr, aclDataType::ACL_FLOAT, max);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87 87 
88 return ACL_SUCCESS;88 return ACL_SUCCESS;
@@ -151,7 +151,7 @@ int main()
151 151 
152 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改152 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
153 auto size = GetShapeSize(selfShape);153 auto size = GetShapeSize(selfShape);
154- std::vector<double> resultData(size, 0);154+ std::vector<float> resultData(size, 0);
155 ret = aclrtMemcpy(155 ret = aclrtMemcpy(
156 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),156 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
157 ACL_MEMCPY_DEVICE_TO_HOST);157 ACL_MEMCPY_DEVICE_TO_HOST);
Mconversion/clip_by_value_v2/examples/test_aclnn_inplace_clamp_min_tensor.cpp+5-5
@@ -76,14 +76,14 @@ int PrepareInputAndOutput(
76 std::vector<int64_t>& selfShape, std::vector<int64_t>& minShape, void** selfDeviceAddr, aclTensor** self,76 std::vector<int64_t>& selfShape, std::vector<int64_t>& minShape, void** selfDeviceAddr, aclTensor** self,
77 void** minDeviceAddr, aclTensor** min)77 void** minDeviceAddr, aclTensor** min)
78{78{
79- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};79+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
80- std::vector<double> minHostData = {2, 1, 1, 2, 2, 6, 6, 9};80+ std::vector<float> minHostData = {2, 1, 1, 2, 2, 6, 6, 9};
81 81 
82 // 创建self aclTensor82 // 创建self aclTensor
83- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);83+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
84 CHECK_RET(ret == ACL_SUCCESS, return ret);84 CHECK_RET(ret == ACL_SUCCESS, return ret);
85 // 创建min aclTensor85 // 创建min aclTensor
86- ret = CreateAclTensor(minHostData, minShape, minDeviceAddr, aclDataType::ACL_DOUBLE, min);86+ ret = CreateAclTensor(minHostData, minShape, minDeviceAddr, aclDataType::ACL_FLOAT, min);
87 CHECK_RET(ret == ACL_SUCCESS, return ret);87 CHECK_RET(ret == ACL_SUCCESS, return ret);
88 88 
89 return ACL_SUCCESS;89 return ACL_SUCCESS;
@@ -151,7 +151,7 @@ int main()
151 151 
152 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧152 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧
153 auto size = GetShapeSize(selfShape);153 auto size = GetShapeSize(selfShape);
154- std::vector<double> resultData(size, 0);154+ std::vector<float> resultData(size, 0);
155 ret = aclrtMemcpy(155 ret = aclrtMemcpy(
156 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),156 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
157 ACL_MEMCPY_DEVICE_TO_HOST);157 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/acos/examples/test_aclnn_acos.cpp+5-5
@@ -85,13 +85,13 @@ int main() {
85 void* outDeviceAddr = nullptr;85 void* outDeviceAddr = nullptr;
86 aclTensor* self = nullptr;86 aclTensor* self = nullptr;
87 aclTensor* out = nullptr;87 aclTensor* out = nullptr;
88- std::vector<double> selfHostData = {1, -1, 0, 0.5, -1.732/2, 12, NAN, -INFINITY};88+ std::vector<float> selfHostData = {1, -1, 0, 0.5, -1.732/2, 12, NAN, -INFINITY};
89- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};89+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
90 // 创建self aclTensor90 // 创建self aclTensor
91- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);91+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
92 CHECK_RET(ret == ACL_SUCCESS, return ret);92 CHECK_RET(ret == ACL_SUCCESS, return ret);
93 // 创建out aclTensor93 // 创建out aclTensor
94- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);94+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 96 
97 uint64_t workspaceSize = 0;97 uint64_t workspaceSize = 0;
@@ -118,7 +118,7 @@ int main() {
118 118 
119 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改119 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
120 auto size = GetShapeSize(outShape);120 auto size = GetShapeSize(outShape);
121- std::vector<double> resultData(size, 0);121+ std::vector<float> resultData(size, 0);
122 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,122 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
123 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);123 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
124 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);124 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/acos/examples/test_aclnn_inplace_acos.cpp+4-4
@@ -83,10 +83,10 @@ int main() {
83 std::vector<int64_t> outShape = {4, 2};83 std::vector<int64_t> outShape = {4, 2};
84 void* selfDeviceAddr = nullptr;84 void* selfDeviceAddr = nullptr;
85 aclTensor* self = nullptr;85 aclTensor* self = nullptr;
86- std::vector<double> selfHostData = {1, -1, 0, 0.5, -1.732/2, 12, NAN, -INFINITY};86+ std::vector<float> selfHostData = {1, -1, 0, 0.5, -1.732/2, 12, NAN, -INFINITY};
87- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};87+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
88 // 创建self aclTensor88 // 创建self aclTensor
89- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);89+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
90 CHECK_RET(ret == ACL_SUCCESS, return ret);90 CHECK_RET(ret == ACL_SUCCESS, return ret);
91 91 
92 uint64_t workspaceSize = 0;92 uint64_t workspaceSize = 0;
@@ -114,7 +114,7 @@ int main() {
114 114 
115 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧115 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧
116 auto size = GetShapeSize(outShape);116 auto size = GetShapeSize(outShape);
117- std::vector<double> resultData(size, 0);117+ std::vector<float> resultData(size, 0);
118 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,118 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
119 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);119 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
120 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);120 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/add/examples/test_aclnn_add.cpp+9-9
@@ -92,21 +92,21 @@ int main()
92 aclTensor* other = nullptr;92 aclTensor* other = nullptr;
93 aclScalar* alpha = nullptr;93 aclScalar* alpha = nullptr;
94 aclTensor* out = nullptr;94 aclTensor* out = nullptr;
95- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};95+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
96- std::vector<double> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};96+ std::vector<float> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};
97- std::vector<double> outHostData(8, 0);97+ std::vector<float> outHostData(8, 0);
98- double alphaValue = 1.2f;98+ float alphaValue = 1.2f;
99 // 创建self aclTensor99 // 创建self aclTensor
100- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);100+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
101 CHECK_RET(ret == ACL_SUCCESS, return ret);101 CHECK_RET(ret == ACL_SUCCESS, return ret);
102 // 创建other aclTensor102 // 创建other aclTensor
103- ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_DOUBLE, &other);103+ ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_FLOAT, &other);
104 CHECK_RET(ret == ACL_SUCCESS, return ret);104 CHECK_RET(ret == ACL_SUCCESS, return ret);
105 // 创建alpha aclScalar105 // 创建alpha aclScalar
106- alpha = aclCreateScalar(&alphaValue, aclDataType::ACL_DOUBLE);106+ alpha = aclCreateScalar(&alphaValue, aclDataType::ACL_FLOAT);
107 CHECK_RET(alpha != nullptr, return ret);107 CHECK_RET(alpha != nullptr, return ret);
108 // 创建out aclTensor108 // 创建out aclTensor
109- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);109+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
110 CHECK_RET(ret == ACL_SUCCESS, return ret);110 CHECK_RET(ret == ACL_SUCCESS, return ret);
111 111 
112 uint64_t workspaceSize = 0;112 uint64_t workspaceSize = 0;
@@ -133,7 +133,7 @@ int main()
133 133 
134 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改134 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
135 auto size = GetShapeSize(outShape);135 auto size = GetShapeSize(outShape);
136- std::vector<double> resultData(size, 0);136+ std::vector<float> resultData(size, 0);
137 ret = aclrtMemcpy(137 ret = aclrtMemcpy(
138 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),138 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),
139 ACL_MEMCPY_DEVICE_TO_HOST);139 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/add/examples/test_aclnn_inplace_add.cpp+7-7
@@ -90,17 +90,17 @@ int main()
90 aclTensor* self = nullptr;90 aclTensor* self = nullptr;
91 aclTensor* other = nullptr;91 aclTensor* other = nullptr;
92 aclScalar* alpha = nullptr;92 aclScalar* alpha = nullptr;
93- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};93+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
94- std::vector<double> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};94+ std::vector<float> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};
95- double alphaValue = 1.2f;95+ float alphaValue = 1.2f;
96 // 创建self aclTensor96 // 创建self aclTensor
97- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);97+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
98 CHECK_RET(ret == ACL_SUCCESS, return ret);98 CHECK_RET(ret == ACL_SUCCESS, return ret);
99 // 创建other aclTensor99 // 创建other aclTensor
100- ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_DOUBLE, &other);100+ ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_FLOAT, &other);
101 CHECK_RET(ret == ACL_SUCCESS, return ret);101 CHECK_RET(ret == ACL_SUCCESS, return ret);
102 // 创建alpha aclScalar102 // 创建alpha aclScalar
103- alpha = aclCreateScalar(&alphaValue, aclDataType::ACL_DOUBLE);103+ alpha = aclCreateScalar(&alphaValue, aclDataType::ACL_FLOAT);
104 CHECK_RET(alpha != nullptr, return ret);104 CHECK_RET(alpha != nullptr, return ret);
105 105 
106 uint64_t workspaceSize = 0;106 uint64_t workspaceSize = 0;
@@ -128,7 +128,7 @@ int main()
128 128 
129 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改129 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
130 auto size = GetShapeSize(outShape);130 auto size = GetShapeSize(outShape);
131- std::vector<double> resultData(size, 0);131+ std::vector<float> resultData(size, 0);
132 ret = aclrtMemcpy(132 ret = aclrtMemcpy(
133 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),133 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
134 ACL_MEMCPY_DEVICE_TO_HOST);134 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/atan/examples/test_aclnn_atan.cpp+5-5
@@ -84,13 +84,13 @@ int main() {
84 void* outDeviceAddr = nullptr;84 void* outDeviceAddr = nullptr;
85 aclTensor* self = nullptr;85 aclTensor* self = nullptr;
86 aclTensor* out = nullptr;86 aclTensor* out = nullptr;
87- std::vector<double> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};87+ std::vector<float> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};
88- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};88+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
89 // 创建self aclTensor89 // 创建self aclTensor
90- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);90+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 // 创建out aclTensor92 // 创建out aclTensor
93- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);93+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
94 CHECK_RET(ret == ACL_SUCCESS, return ret);94 CHECK_RET(ret == ACL_SUCCESS, return ret);
95 95 
96 // 3. aclnnAtan接口调用示例96 // 3. aclnnAtan接口调用示例
@@ -115,7 +115,7 @@ int main() {
115 115 
116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
117 auto size = GetShapeSize(outShape);117 auto size = GetShapeSize(outShape);
118- std::vector<double> resultData(size, 0);118+ std::vector<float> resultData(size, 0);
119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/atan/examples/test_aclnn_inplace_atan.cpp+4-4
@@ -82,10 +82,10 @@ int main() {
82 std::vector<int64_t> outShape = {4, 2};82 std::vector<int64_t> outShape = {4, 2};
83 void* selfDeviceAddr = nullptr;83 void* selfDeviceAddr = nullptr;
84 aclTensor* self = nullptr;84 aclTensor* self = nullptr;
85- std::vector<double> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};85+ std::vector<float> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};
86- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};86+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
87 // 创建self aclTensor87 // 创建self aclTensor
88- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);88+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
89 CHECK_RET(ret == ACL_SUCCESS, return ret);89 CHECK_RET(ret == ACL_SUCCESS, return ret);
90 90 
91 // 3. aclnnInplaceAtan接口调用示例91 // 3. aclnnInplaceAtan接口调用示例
@@ -110,7 +110,7 @@ int main() {
110 110 
111 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改111 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
112 auto size = GetShapeSize(outShape);112 auto size = GetShapeSize(outShape);
113- std::vector<double> resultData(size, 0);113+ std::vector<float> resultData(size, 0);
114 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,114 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
115 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);115 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
116 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);116 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/cast/examples/test_aclnn_cast.cpp+6-6
@@ -84,21 +84,21 @@ int main() {
84 aclTensor* self = nullptr;84 aclTensor* self = nullptr;
85 aclTensor* out = nullptr;85 aclTensor* out = nullptr;
86 86 
87- std::vector<float> selfHostData = {0.1, 1.1, 2.1, 3.1, 4.1, 5.1, 6.1, 7.1};87+ std::vector<double> selfHostData = {0.1, 1.1, 2.1, 3.1, 4.1, 5.1, 6.1, 7.1};
88- std::vector<double> outHostData = {0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0};88+ std::vector<float> outHostData = {0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0};
89 89 
90 // 创建self aclTensor90 // 创建self aclTensor
91- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);91+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);
92 CHECK_RET(ret == ACL_SUCCESS, return ret);92 CHECK_RET(ret == ACL_SUCCESS, return ret);
93 // 创建out aclTensor93 // 创建out aclTensor
94- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);94+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 96 
97 // 3. 调用CANN算子库API，需要修改为具体的Api名称97 // 3. 调用CANN算子库API，需要修改为具体的Api名称
98 uint64_t workspaceSize = 0;98 uint64_t workspaceSize = 0;
99 aclOpExecutor* executor;99 aclOpExecutor* executor;
100 // 调用aclnnCast第一段接口100 // 调用aclnnCast第一段接口
101- ret = aclnnCastGetWorkspaceSize(self, aclDataType::ACL_DOUBLE, out, &workspaceSize, &executor);101+ ret = aclnnCastGetWorkspaceSize(self, aclDataType::ACL_FLOAT, out, &workspaceSize, &executor);
102 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnCastGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);102 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnCastGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
103 // 根据第一段接口计算出的workspaceSize申请device内存103 // 根据第一段接口计算出的workspaceSize申请device内存
104 void* workspaceAddr = nullptr;104 void* workspaceAddr = nullptr;
@@ -116,7 +116,7 @@ int main() {
116 116 
117 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改117 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
118 auto size = GetShapeSize(outShape);118 auto size = GetShapeSize(outShape);
119- std::vector<double> resultData(size, 0);119+ std::vector<float> resultData(size, 0);
120 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,120 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
121 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);121 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
122 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);122 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/ceil/examples/test_aclnn_ceil.cpp+5-5
@@ -83,15 +83,15 @@ aclError CreateInputs(
83 std::vector<int64_t>& inputShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, void** outDeviceAddr,83 std::vector<int64_t>& inputShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, void** outDeviceAddr,
84 aclTensor** self, aclTensor** out)84 aclTensor** self, aclTensor** out)
85{85{
86- std::vector<double> selfHostData = {0, 1.1, 2.2, 3.3, 4.4, 5.5, 6.6, 7.7};86+ std::vector<float> selfHostData = {0, 1.1, 2.2, 3.3, 4.4, 5.5, 6.6, 7.7};
87- std::vector<double> outHostData(8, 0);87+ std::vector<float> outHostData(8, 0);
88 88 
89 // 创建 self aclTensor89 // 创建 self aclTensor
90- auto ret = CreateAclTensor(selfHostData, inputShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);90+ auto ret = CreateAclTensor(selfHostData, inputShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 92 
93 // 创建 out aclTensor93 // 创建 out aclTensor
94- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);94+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 96 
97 return ACL_SUCCESS;97 return ACL_SUCCESS;
@@ -103,7 +103,7 @@ aclError ExecOpApi(
103{103{
104 aclOpExecutor* executor;104 aclOpExecutor* executor;
105 auto size = GetShapeSize(outShape);105 auto size = GetShapeSize(outShape);
106- std::vector<double> resultData(size, 0);106+ std::vector<float> resultData(size, 0);
107 107 
108 // aclnnCeil 接口调用示例108 // aclnnCeil 接口调用示例
109 LOG_PRINT("test aclnnCeil\n");109 LOG_PRINT("test aclnnCeil\n");
Mmath/cross/examples/test_aclnn_linalg_cross.cpp+7-7
@@ -84,20 +84,20 @@ aclError CreateInputs(
84 void** selfDeviceAddr, void** otherDeviceAddr, void** outDeviceAddr, aclTensor** self, aclTensor** other,84 void** selfDeviceAddr, void** otherDeviceAddr, void** outDeviceAddr, aclTensor** self, aclTensor** other,
85 aclTensor** out)85 aclTensor** out)
86{86{
87- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7, 8};87+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7, 8};
88- std::vector<double> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3, 3};88+ std::vector<float> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3, 3};
89- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0, 0};89+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0, 0};
90 90 
91 // 创建 self aclTensor91 // 创建 self aclTensor
92- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);92+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 94 
95 // 创建 other aclTensor95 // 创建 other aclTensor
96- ret = CreateAclTensor(otherHostData, otherShape, otherDeviceAddr, aclDataType::ACL_DOUBLE, other);96+ ret = CreateAclTensor(otherHostData, otherShape, otherDeviceAddr, aclDataType::ACL_FLOAT, other);
97 CHECK_RET(ret == ACL_SUCCESS, return ret);97 CHECK_RET(ret == ACL_SUCCESS, return ret);
98 98 
99 // 创建 out aclTensor99 // 创建 out aclTensor
100- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);100+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
101 CHECK_RET(ret == ACL_SUCCESS, return ret);101 CHECK_RET(ret == ACL_SUCCESS, return ret);
102 102 
103 return ACL_SUCCESS;103 return ACL_SUCCESS;
@@ -131,7 +131,7 @@ aclError ExecOpApi(
131 131 
132 // 从 device 拷贝结果到 host132 // 从 device 拷贝结果到 host
133 auto size = GetShapeSize(outShape);133 auto size = GetShapeSize(outShape);
134- std::vector<double> resultData(size, 0);134+ std::vector<float> resultData(size, 0);
135 135 
136 ret = aclrtMemcpy(136 ret = aclrtMemcpy(
137 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),137 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),
Mmath/equal/examples/test_aclnn_inplace_eq_scalar.cpp+5-5
@@ -81,15 +81,15 @@ aclError InitAcl(int32_t deviceId, aclrtStream* stream)
81 81 
82aclError CreateInputs(std::vector<int64_t>& selfShape, void** selfDeviceAddr, aclTensor** self, aclScalar** other)82aclError CreateInputs(std::vector<int64_t>& selfShape, void** selfDeviceAddr, aclTensor** self, aclScalar** other)
83{83{
84- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};84+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
85- double otherValue = 2.0;85+ float otherValue = 2.0;
86 86 
87 // 创建self aclTensor87 // 创建self aclTensor
88- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);88+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
89 CHECK_RET(ret == ACL_SUCCESS, return ret);89 CHECK_RET(ret == ACL_SUCCESS, return ret);
90 90 
91 // 创建other aclScalar91 // 创建other aclScalar
92- *other = aclCreateScalar(&otherValue, aclDataType::ACL_DOUBLE);92+ *other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
93 CHECK_RET(*other != nullptr, return ACL_ERROR_INVALID_PARAM);93 CHECK_RET(*other != nullptr, return ACL_ERROR_INVALID_PARAM);
94 94 
95 return ACL_SUCCESS;95 return ACL_SUCCESS;
@@ -124,7 +124,7 @@ aclError ExecOpApi(
124 124 
125 // 拷贝输出125 // 拷贝输出
126 auto size = GetShapeSize(selfShape);126 auto size = GetShapeSize(selfShape);
127- std::vector<double> resultData(size);127+ std::vector<float> resultData(size);
128 128 
129 ret = aclrtMemcpy(129 ret = aclrtMemcpy(
130 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),130 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
Mmath/equal/examples/test_aclnn_inplace_eq_tensor.cpp+6-6
@@ -83,15 +83,15 @@ aclError CreateInputs(
83 std::vector<int64_t>& selfShape, std::vector<int64_t>& otherShape, void** selfDeviceAddr, void** otherDeviceAddr,83 std::vector<int64_t>& selfShape, std::vector<int64_t>& otherShape, void** selfDeviceAddr, void** otherDeviceAddr,
84 aclTensor** self, aclTensor** other)84 aclTensor** self, aclTensor** other)
85{85{
86- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};86+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
87- std::vector<double> otherHostData = {1, 1, 3, 3, 5, 5, 7, 7};87+ std::vector<float> otherHostData = {1, 1, 3, 3, 5, 5, 7, 7};
88 88 
89 // 创建 self89 // 创建 self
90- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);90+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 92 
93 // 创建 other93 // 创建 other
94- ret = CreateAclTensor(otherHostData, otherShape, otherDeviceAddr, aclDataType::ACL_DOUBLE, other);94+ ret = CreateAclTensor(otherHostData, otherShape, otherDeviceAddr, aclDataType::ACL_FLOAT, other);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 96 
97 return ACL_SUCCESS;97 return ACL_SUCCESS;
@@ -125,10 +125,10 @@ aclError ExecOpApi(
125 125 
126 // 拷贝输出126 // 拷贝输出
127 auto size = GetShapeSize(selfShape);127 auto size = GetShapeSize(selfShape);
128- std::vector<double> resultData(size);128+ std::vector<float> resultData(size);
129 129 
130 ret = aclrtMemcpy(130 ret = aclrtMemcpy(
131- resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(double),131+ resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
132 ACL_MEMCPY_DEVICE_TO_HOST);132 ACL_MEMCPY_DEVICE_TO_HOST);
133 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);133 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
134 134 
Mmath/greater/examples/test_aclnn_inplace_gt_scalar.cpp+5-5
@@ -79,13 +79,13 @@ int main() {
79 void* selfDeviceAddr = nullptr;79 void* selfDeviceAddr = nullptr;
80 aclTensor* self = nullptr;80 aclTensor* self = nullptr;
81 aclScalar* other = nullptr;81 aclScalar* other = nullptr;
82- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};82+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
83- double otherValue = 3.5;83+ float otherValue = 3.5;
84 // 创建self aclTensor84 // 创建self aclTensor
85- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);85+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87 // 创建other aclScalar87 // 创建other aclScalar
88- other = aclCreateScalar(&otherValue, aclDataType::ACL_DOUBLE);88+ other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
89 CHECK_RET(other != nullptr, return ret);89 CHECK_RET(other != nullptr, return ret);
90 90 
91 // 3. 调用CANN算子库API，需要修改为具体的Api名称91 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -110,7 +110,7 @@ int main() {
110 110 
111 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改111 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
112 auto size = GetShapeSize(selfShape);112 auto size = GetShapeSize(selfShape);
113- std::vector<double> resultData(size, 0);113+ std::vector<float> resultData(size, 0);
114 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,114 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
115 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);115 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
116 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);116 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/greater/examples/test_aclnn_inplace_gt_tensor.cpp+5-5
@@ -82,14 +82,14 @@ int main() {
82 void* otherDeviceAddr = nullptr;82 void* otherDeviceAddr = nullptr;
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclTensor* other = nullptr;84 aclTensor* other = nullptr;
85- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};85+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
86- std::vector<double> otherHostData = {0, 1, 1, 2, 3, 4, 5, 6};86+ std::vector<float> otherHostData = {0, 1, 1, 2, 3, 4, 5, 6};
87 87 
88 // 创建self aclTensor88 // 创建self aclTensor
89- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);89+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
90 CHECK_RET(ret == ACL_SUCCESS, return ret);90 CHECK_RET(ret == ACL_SUCCESS, return ret);
91 // 创建other aclTensor91 // 创建other aclTensor
92- ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_DOUBLE, &other);92+ ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_FLOAT, &other);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 94 
95 // 3. 调用CANN算子库API，需要修改为具体的API95 // 3. 调用CANN算子库API，需要修改为具体的API
@@ -112,7 +112,7 @@ int main() {
112 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);112 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
113 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改113 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
114 auto size = GetShapeSize(selfShape);114 auto size = GetShapeSize(selfShape);
115- std::vector<double> resultData(size, 0);115+ std::vector<float> resultData(size, 0);
116 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),116 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
117 ACL_MEMCPY_DEVICE_TO_HOST);117 ACL_MEMCPY_DEVICE_TO_HOST);
118 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);118 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/less/examples/test_aclnn_inplace_lt_scalar.cpp+5-5
@@ -79,12 +79,12 @@ int ExecuteInplaceLtScalarOperator(aclrtStream stream)
79 void* selfDeviceAddr = nullptr;79 void* selfDeviceAddr = nullptr;
80 aclTensor* self = nullptr;80 aclTensor* self = nullptr;
81 aclScalar* other = nullptr;81 aclScalar* other = nullptr;
82- std::vector<double> selfHostData = {0, 1, 1.2, 0.3, 4.1, 5, 1.6, 7};82+ std::vector<float> selfHostData = {0, 1, 1.2, 0.3, 4.1, 5, 1.6, 7};
83- double otherValue = 1.2;83+ float otherValue = 1.2;
84 84 
85- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);85+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
86 CHECK_RET(ret == ACL_SUCCESS, return ret);86 CHECK_RET(ret == ACL_SUCCESS, return ret);
87- other = aclCreateScalar(&otherValue, aclDataType::ACL_DOUBLE);87+ other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
88 CHECK_RET(other != nullptr, return ret);88 CHECK_RET(other != nullptr, return ret);
89 89 
90 uint64_t workspaceSize = 0;90 uint64_t workspaceSize = 0;
@@ -106,7 +106,7 @@ int ExecuteInplaceLtScalarOperator(aclrtStream stream)
106 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);106 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
107 107 
108 auto size = GetShapeSize(selfShape);108 auto size = GetShapeSize(selfShape);
109- std::vector<double> resultData(size, 0);109+ std::vector<float> resultData(size, 0);
110 ret = aclrtMemcpy(110 ret = aclrtMemcpy(
111 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),111 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
112 ACL_MEMCPY_DEVICE_TO_HOST);112 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/less/examples/test_aclnn_inplace_lt_tensor.cpp+5-5
@@ -81,10 +81,10 @@ int ExecuteInplaceLtTensorOperator(aclrtStream stream)
81 void* otherDeviceAddr = nullptr;81 void* otherDeviceAddr = nullptr;
82 aclTensor* self = nullptr;82 aclTensor* self = nullptr;
83 aclTensor* other = nullptr;83 aclTensor* other = nullptr;
84- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};84+ std::vector<int32_t> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
85- std::vector<int> otherHostData = {1, 1, 1, 1, 0, 0, 0, 0};85+ std::vector<int32_t> otherHostData = {1, 1, 1, 1, 0, 0, 0, 0};
86 86 
87- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);87+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_INT32, &self);
88 CHECK_RET(ret == ACL_SUCCESS, return ret);88 CHECK_RET(ret == ACL_SUCCESS, return ret);
89 ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_INT32, &other);89 ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_INT32, &other);
90 CHECK_RET(ret == ACL_SUCCESS, return ret);90 CHECK_RET(ret == ACL_SUCCESS, return ret);
@@ -108,13 +108,13 @@ int ExecuteInplaceLtTensorOperator(aclrtStream stream)
108 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);108 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
109 109 
110 auto size = GetShapeSize(selfShape);110 auto size = GetShapeSize(selfShape);
111- std::vector<double> resultData(size, 0);111+ std::vector<int32_t> resultData(size, 0);
112 ret = aclrtMemcpy(112 ret = aclrtMemcpy(
113 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),113 resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, size * sizeof(resultData[0]),
114 ACL_MEMCPY_DEVICE_TO_HOST);114 ACL_MEMCPY_DEVICE_TO_HOST);
115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
116 for (int64_t i = 0; i < size; i++) {116 for (int64_t i = 0; i < size; i++) {
117- LOG_PRINT("result[%ld] is: %lf\n", i, resultData[i]);117+ LOG_PRINT("result[%ld] is: %d\n", i, resultData[i]);
118 }118 }
119 119 
120 aclDestroyTensor(self);120 aclDestroyTensor(self);
Mmath/less/examples/test_aclnn_lt_scalar.cpp+7-7
@@ -83,17 +83,17 @@ int ExecuteLtScalarOperator(aclrtStream stream)
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclScalar* other = nullptr;84 aclScalar* other = nullptr;
85 aclTensor* out = nullptr;85 aclTensor* out = nullptr;
86- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};86+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
87- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};87+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
88- double otherValue = 1.2;88+ float otherValue = 1.2;
89 // 创建self aclTensor89 // 创建self aclTensor
90- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);90+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 // 创建other aclScalar92 // 创建other aclScalar
93- other = aclCreateScalar(&otherValue, aclDataType::ACL_DOUBLE);93+ other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
94 CHECK_RET(other != nullptr, return ret);94 CHECK_RET(other != nullptr, return ret);
95 // 创建out aclTensor95 // 创建out aclTensor
96- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);96+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
97 CHECK_RET(ret == ACL_SUCCESS, return ret);97 CHECK_RET(ret == ACL_SUCCESS, return ret);
98 98 
99 // 3. 调用CANN算子库API，需要修改为具体的Api名称99 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -118,7 +118,7 @@ int ExecuteLtScalarOperator(aclrtStream stream)
118 118 
119 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改119 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
120 auto size = GetShapeSize(outShape);120 auto size = GetShapeSize(outShape);
121- std::vector<double> resultData(size, 0);121+ std::vector<float> resultData(size, 0);
122 ret = aclrtMemcpy(122 ret = aclrtMemcpy(
123 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),123 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),
124 ACL_MEMCPY_DEVICE_TO_HOST);124 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/less/examples/test_aclnn_lt_tensor.cpp+7-7
@@ -82,9 +82,9 @@ struct LtTensorData {
82 aclTensor* self = nullptr;82 aclTensor* self = nullptr;
83 aclTensor* other = nullptr;83 aclTensor* other = nullptr;
84 aclTensor* out = nullptr;84 aclTensor* out = nullptr;
85- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};85+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
86- std::vector<double> otherHostData = {5, 5, 5, 5, 5, 5, 5, 5};86+ std::vector<float> otherHostData = {5, 5, 5, 5, 5, 5, 5, 5};
87- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};87+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
88 void* workspaceAddr = nullptr;88 void* workspaceAddr = nullptr;
89 uint64_t workspaceSize = 0;89 uint64_t workspaceSize = 0;
90};90};
@@ -94,14 +94,14 @@ int CreateInputAndOutputTensors(LtTensorData& data)
94 auto ret = 0;94 auto ret = 0;
95 95 
96 // 创建self aclTensor96 // 创建self aclTensor
97- ret = CreateAclTensor(data.selfHostData, data.selfShape, &data.selfDeviceAddr, aclDataType::ACL_DOUBLE, &data.self);97+ ret = CreateAclTensor(data.selfHostData, data.selfShape, &data.selfDeviceAddr, aclDataType::ACL_FLOAT, &data.self);
98 CHECK_RET(ret == ACL_SUCCESS, return ret);98 CHECK_RET(ret == ACL_SUCCESS, return ret);
99 // 创建other aclTensor99 // 创建other aclTensor
100 ret = CreateAclTensor(100 ret = CreateAclTensor(
101- data.otherHostData, data.otherShape, &data.otherDeviceAddr, aclDataType::ACL_DOUBLE, &data.other);101+ data.otherHostData, data.otherShape, &data.otherDeviceAddr, aclDataType::ACL_FLOAT, &data.other);
102 CHECK_RET(ret == ACL_SUCCESS, return ret);102 CHECK_RET(ret == ACL_SUCCESS, return ret);
103 // 创建out aclTensor103 // 创建out aclTensor
104- ret = CreateAclTensor(data.outHostData, data.outShape, &data.outDeviceAddr, aclDataType::ACL_DOUBLE, &data.out);104+ ret = CreateAclTensor(data.outHostData, data.outShape, &data.outDeviceAddr, aclDataType::ACL_FLOAT, &data.out);
105 CHECK_RET(ret == ACL_SUCCESS, return ret);105 CHECK_RET(ret == ACL_SUCCESS, return ret);
106 106 
107 return ret;107 return ret;
@@ -138,7 +138,7 @@ int ProcessAndPrintResults(const LtTensorData& data)
138{138{
139 auto ret = 0;139 auto ret = 0;
140 auto size = GetShapeSize(data.outShape);140 auto size = GetShapeSize(data.outShape);
141- std::vector<double> resultData(size, 0);141+ std::vector<float> resultData(size, 0);
142 ret = aclrtMemcpy(142 ret = aclrtMemcpy(
143 resultData.data(), resultData.size() * sizeof(resultData[0]), data.outDeviceAddr, size * sizeof(resultData[0]),143 resultData.data(), resultData.size() * sizeof(resultData[0]), data.outDeviceAddr, size * sizeof(resultData[0]),
144 ACL_MEMCPY_DEVICE_TO_HOST);144 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/less/examples/test_aclnn_signbit.cpp+5-5
@@ -88,13 +88,13 @@ int main()
88 void* outDeviceAddr = nullptr;88 void* outDeviceAddr = nullptr;
89 aclTensor* self = nullptr;89 aclTensor* self = nullptr;
90 aclTensor* out = nullptr;90 aclTensor* out = nullptr;
91- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};91+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
92- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};92+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
93 // 创建self aclTensor93 // 创建self aclTensor
94- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);94+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
95 CHECK_RET(ret == ACL_SUCCESS, return ret);95 CHECK_RET(ret == ACL_SUCCESS, return ret);
96 // 创建out aclTensor96 // 创建out aclTensor
97- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);97+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
98 CHECK_RET(ret == ACL_SUCCESS, return ret);98 CHECK_RET(ret == ACL_SUCCESS, return ret);
99 99 
100 // 3. 调用CANN算子库API100 // 3. 调用CANN算子库API
@@ -119,7 +119,7 @@ int main()
119 119 
120 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改120 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
121 auto size = GetShapeSize(outShape);121 auto size = GetShapeSize(outShape);
122- std::vector<double> resultData(size, 0);122+ std::vector<float> resultData(size, 0);
123 ret = aclrtMemcpy(123 ret = aclrtMemcpy(
124 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),124 resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(resultData[0]),
125 ACL_MEMCPY_DEVICE_TO_HOST);125 ACL_MEMCPY_DEVICE_TO_HOST);
Mmath/log/examples/test_aclnn_inplace_log.cpp+3-3
@@ -82,9 +82,9 @@ int main() {
82 void* outDeviceAddr = nullptr;82 void* outDeviceAddr = nullptr;
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclTensor* out = nullptr;84 aclTensor* out = nullptr;
85- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};85+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
86 // 创建self aclTensor86 // 创建self aclTensor
87- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);87+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
88 CHECK_RET(ret == ACL_SUCCESS, return ret);88 CHECK_RET(ret == ACL_SUCCESS, return ret);
89 89 
90 // 3. 调用CANN算子库API，需要修改为具体的Api名称90 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -109,7 +109,7 @@ int main() {
109 109 
110 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改110 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
111 auto size = GetShapeSize(selfShape);111 auto size = GetShapeSize(selfShape);
112- std::vector<double> resultData(size, 0);112+ std::vector<float> resultData(size, 0);
113 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,113 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
114 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);114 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/log/examples/test_aclnn_inplace_log10.cpp+5-5
@@ -83,14 +83,14 @@ int main() {
83 void* outDeviceAddr = nullptr;83 void* outDeviceAddr = nullptr;
84 aclTensor* self = nullptr;84 aclTensor* self = nullptr;
85 aclTensor* out = nullptr;85 aclTensor* out = nullptr;
86- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};86+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
87- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};87+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
88 88 
89 // 创建self aclTensor89 // 创建self aclTensor
90- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);90+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 // 创建out aclTensor92 // 创建out aclTensor
93- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);93+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
94 CHECK_RET(ret == ACL_SUCCESS, return ret);94 CHECK_RET(ret == ACL_SUCCESS, return ret);
95 95 
96 // 3. 调用CANN算子库API96 // 3. 调用CANN算子库API
@@ -122,7 +122,7 @@ int main() {
122 // 5. 将device侧内存上的结果拷贝到host侧122 // 5. 将device侧内存上的结果拷贝到host侧
123 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改123 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
124 auto size = GetShapeSize(outShape);124 auto size = GetShapeSize(outShape);
125- std::vector<double> resultData(size, 0); 125+ std::vector<float> resultData(size, 0);
126 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,126 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
127 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);127 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
128 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);128 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/log/examples/test_aclnn_inplace_log2.cpp+3-3
@@ -78,9 +78,9 @@ int main() {
78 std::vector<int64_t> selfShape = {4, 2};78 std::vector<int64_t> selfShape = {4, 2};
79 void* selfDeviceAddr = nullptr;79 void* selfDeviceAddr = nullptr;
80 aclTensor* selfRef = nullptr;80 aclTensor* selfRef = nullptr;
81- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 8};81+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 8};
82 // 创建self aclTensor82 // 创建self aclTensor
83- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &selfRef);83+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &selfRef);
84 CHECK_RET(ret == ACL_SUCCESS, return ret);84 CHECK_RET(ret == ACL_SUCCESS, return ret);
85 85 
86 // 3. 调用CANN算子库API，需要修改为具体的Api名称86 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -103,7 +103,7 @@ int main() {
103 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);103 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
104 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改104 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
105 auto size = GetShapeSize(selfShape);105 auto size = GetShapeSize(selfShape);
106- std::vector<double> resultData(size, 0);106+ std::vector<float> resultData(size, 0);
107 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,107 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
108 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);108 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
109 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);109 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/log/examples/test_aclnn_log.cpp+5-5
@@ -82,13 +82,13 @@ int main() {
82 void* outDeviceAddr = nullptr;82 void* outDeviceAddr = nullptr;
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclTensor* out = nullptr;84 aclTensor* out = nullptr;
85- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};85+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
86- std::vector<double> outHostData(8, 0);86+ std::vector<float> outHostData(8, 0);
87 // 创建self aclTensor87 // 创建self aclTensor
88- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);88+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
89 CHECK_RET(ret == ACL_SUCCESS, return ret);89 CHECK_RET(ret == ACL_SUCCESS, return ret);
90 // 创建out aclTensor90 // 创建out aclTensor
91- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);91+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
92 CHECK_RET(ret == ACL_SUCCESS, return ret);92 CHECK_RET(ret == ACL_SUCCESS, return ret);
93 93 
94 // 3. 调用CANN算子库API，需要修改为具体的Api名称94 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -113,7 +113,7 @@ int main() {
113 113 
114 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改114 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
115 auto size = GetShapeSize(outShape);115 auto size = GetShapeSize(outShape);
116- std::vector<double> resultData(size, 0);116+ std::vector<float> resultData(size, 0);
117 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,117 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
118 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);118 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
119 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);119 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/log/examples/test_aclnn_log10.cpp+5-5
@@ -83,14 +83,14 @@ int main() {
83 void* outDeviceAddr = nullptr;83 void* outDeviceAddr = nullptr;
84 aclTensor* self = nullptr;84 aclTensor* self = nullptr;
85 aclTensor* out = nullptr;85 aclTensor* out = nullptr;
86- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};86+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
87- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};87+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
88 88 
89 // 创建self aclTensor89 // 创建self aclTensor
90- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);90+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 // 创建out aclTensor92 // 创建out aclTensor
93- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);93+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
94 CHECK_RET(ret == ACL_SUCCESS, return ret);94 CHECK_RET(ret == ACL_SUCCESS, return ret);
95 95 
96 // aclnnLog10接口调用示例96 // aclnnLog10接口调用示例
@@ -114,7 +114,7 @@ int main() {
114 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);114 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
115 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改115 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
116 auto size = GetShapeSize(outShape);116 auto size = GetShapeSize(outShape);
117- std::vector<double> resultData(size, 0);117+ std::vector<float> resultData(size, 0);
118 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,118 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
119 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);119 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
120 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);120 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/log/examples/test_aclnn_log2.cpp+5-5
@@ -82,13 +82,13 @@ int main() {
82 void* outDeviceAddr = nullptr;82 void* outDeviceAddr = nullptr;
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclTensor* out = nullptr;84 aclTensor* out = nullptr;
85- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};85+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
86- std::vector<double> outHostData(8, 0);86+ std::vector<float> outHostData(8, 0);
87 // 创建self aclTensor87 // 创建self aclTensor
88- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);88+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
89 CHECK_RET(ret == ACL_SUCCESS, return ret);89 CHECK_RET(ret == ACL_SUCCESS, return ret);
90 // 创建out aclTensor90 // 创建out aclTensor
91- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);91+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
92 CHECK_RET(ret == ACL_SUCCESS, return ret);92 CHECK_RET(ret == ACL_SUCCESS, return ret);
93 93 
94 // 3. 调用CANN算子库API，需要修改为具体的Api名称94 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -113,7 +113,7 @@ int main() {
113 113 
114 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改114 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
115 auto size = GetShapeSize(outShape);115 auto size = GetShapeSize(outShape);
116- std::vector<double> resultData(size, 0);116+ std::vector<float> resultData(size, 0);
117 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,117 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
118 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);118 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
119 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);119 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/not_equal/examples/test_aclnn_inplace_ne_scalar.cpp+7-7
@@ -95,18 +95,18 @@ int main()
95 aclTensor *self = nullptr;95 aclTensor *self = nullptr;
96 aclScalar *other = nullptr;96 aclScalar *other = nullptr;
97 aclTensor *out = nullptr;97 aclTensor *out = nullptr;
98- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};98+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
99- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};99+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
100- double otherValue = 1.0f;100+ float otherValue = 1.0f;
101 101 
102 // 创建self aclTensor102 // 创建self aclTensor
103- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);103+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
104 CHECK_RET(ret == ACL_SUCCESS, return ret);104 CHECK_RET(ret == ACL_SUCCESS, return ret);
105 // 创建other aclScalar105 // 创建other aclScalar
106- other = aclCreateScalar(&otherValue, aclDataType::ACL_DOUBLE);106+ other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
107 CHECK_RET(ret == ACL_SUCCESS, return ret);107 CHECK_RET(ret == ACL_SUCCESS, return ret);
108 // 创建out aclTensor108 // 创建out aclTensor
109- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);109+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
110 CHECK_RET(ret == ACL_SUCCESS, return ret);110 CHECK_RET(ret == ACL_SUCCESS, return ret);
111 111 
112 // aclnnInplaceNeScalar调用示例112 // aclnnInplaceNeScalar调用示例
@@ -130,7 +130,7 @@ int main()
130 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);130 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
131 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改131 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
132 auto size = GetShapeSize(selfShape);132 auto size = GetShapeSize(selfShape);
133- std::vector<double> resultData(size, 0);133+ std::vector<float> resultData(size, 0);
134 ret = aclrtMemcpy(resultData.data(),134 ret = aclrtMemcpy(resultData.data(),
135 resultData.size() * sizeof(resultData[0]),135 resultData.size() * sizeof(resultData[0]),
136 selfDeviceAddr,136 selfDeviceAddr,
Mmath/not_equal/examples/test_aclnn_inplace_ne_tensor.cpp+7-7
@@ -96,18 +96,18 @@ int main()
96 aclTensor *self = nullptr;96 aclTensor *self = nullptr;
97 aclTensor *other = nullptr;97 aclTensor *other = nullptr;
98 aclTensor *out = nullptr;98 aclTensor *out = nullptr;
99- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};99+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
100- std::vector<double> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};100+ std::vector<float> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};
101- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};101+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
102 102 
103 // 创建self aclTensor103 // 创建self aclTensor
104- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);104+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
105 CHECK_RET(ret == ACL_SUCCESS, return ret);105 CHECK_RET(ret == ACL_SUCCESS, return ret);
106 // 创建other aclTensor106 // 创建other aclTensor
107- ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_DOUBLE, &other);107+ ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_FLOAT, &other);
108 CHECK_RET(ret == ACL_SUCCESS, return ret);108 CHECK_RET(ret == ACL_SUCCESS, return ret);
109 // 创建out aclTensor109 // 创建out aclTensor
110- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);110+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
111 CHECK_RET(ret == ACL_SUCCESS, return ret);111 CHECK_RET(ret == ACL_SUCCESS, return ret);
112 112
113 //aclnnInplaceNeTensor接口调用示例113 //aclnnInplaceNeTensor接口调用示例
@@ -133,7 +133,7 @@ int main()
133 133
134 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改134 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
135 auto size = GetShapeSize(outShape);135 auto size = GetShapeSize(outShape);
136- std::vector<double> resultData(size, 0);136+ std::vector<float> resultData(size, 0);
137 ret = aclrtMemcpy(resultData.data(),137 ret = aclrtMemcpy(resultData.data(),
138 resultData.size() * sizeof(resultData[0]), selfDeviceAddr,138 resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
139 size * sizeof(resultData[0]),139 size * sizeof(resultData[0]),
Mmath/not_equal/examples/test_aclnn_ne_scalar.cpp+7-7
@@ -95,18 +95,18 @@ int main()
95 aclTensor *self = nullptr;95 aclTensor *self = nullptr;
96 aclScalar *other = nullptr;96 aclScalar *other = nullptr;
97 aclTensor *out = nullptr;97 aclTensor *out = nullptr;
98- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};98+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
99- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};99+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
100- double otherValue = 1.0f;100+ float otherValue = 1.0f;
101 101 
102 // 创建self aclTensor102 // 创建self aclTensor
103- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);103+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
104 CHECK_RET(ret == ACL_SUCCESS, return ret);104 CHECK_RET(ret == ACL_SUCCESS, return ret);
105 // 创建other aclScalar105 // 创建other aclScalar
106- other = aclCreateScalar(&otherValue, aclDataType::ACL_DOUBLE);106+ other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
107 CHECK_RET(ret == ACL_SUCCESS, return ret);107 CHECK_RET(ret == ACL_SUCCESS, return ret);
108 // 创建out aclTensor108 // 创建out aclTensor
109- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);109+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
110 CHECK_RET(ret == ACL_SUCCESS, return ret);110 CHECK_RET(ret == ACL_SUCCESS, return ret);
111 111
112 // aclnnNeScalar调用示例112 // aclnnNeScalar调用示例
@@ -130,7 +130,7 @@ int main()
130 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);130 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
131 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改131 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
132 auto size = GetShapeSize(selfShape);132 auto size = GetShapeSize(selfShape);
133- std::vector<double> resultData(size, 0);133+ std::vector<float> resultData(size, 0);
134 ret = aclrtMemcpy(resultData.data(),134 ret = aclrtMemcpy(resultData.data(),
135 resultData.size() * sizeof(resultData[0]),135 resultData.size() * sizeof(resultData[0]),
136 outDeviceAddr,136 outDeviceAddr,
Mmath/not_equal/examples/test_aclnn_ne_tensor.cpp+7-7
@@ -96,18 +96,18 @@ int main()
96 aclTensor *self = nullptr;96 aclTensor *self = nullptr;
97 aclTensor *other = nullptr;97 aclTensor *other = nullptr;
98 aclTensor *out = nullptr;98 aclTensor *out = nullptr;
99- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};99+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
100- std::vector<double> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};100+ std::vector<float> otherHostData = {1, 1, 1, 2, 2, 2, 3, 3};
101- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};101+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
102 102 
103 // 创建self aclTensor103 // 创建self aclTensor
104- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);104+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
105 CHECK_RET(ret == ACL_SUCCESS, return ret);105 CHECK_RET(ret == ACL_SUCCESS, return ret);
106 // 创建other aclTensor106 // 创建other aclTensor
107- ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_DOUBLE, &other);107+ ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_FLOAT, &other);
108 CHECK_RET(ret == ACL_SUCCESS, return ret);108 CHECK_RET(ret == ACL_SUCCESS, return ret);
109 // 创建out aclTensor109 // 创建out aclTensor
110- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);110+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
111 CHECK_RET(ret == ACL_SUCCESS, return ret);111 CHECK_RET(ret == ACL_SUCCESS, return ret);
112 112 
113 // aclnnNeTensor接口调用示例113 // aclnnNeTensor接口调用示例
@@ -133,7 +133,7 @@ int main()
133 133 
134 // 5. 获取输出的值，将Device侧内存上的结果拷贝至Host侧134 // 5. 获取输出的值，将Device侧内存上的结果拷贝至Host侧
135 auto size = GetShapeSize(outShape);135 auto size = GetShapeSize(outShape);
136- std::vector<double> resultData(size, 0);136+ std::vector<float> resultData(size, 0);
137 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,137 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
138 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);138 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
139 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);139 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/pow/op_host/arch35/pow_tensor_scalar_tiling_arch35.cpp+1-1
@@ -143,5 +143,5 @@ ge::graphStatus PowTensorScalarTiling::PostTiling()
143 return ge::GRAPH_SUCCESS;143 return ge::GRAPH_SUCCESS;
144}144}
145 145 
146-REGISTER_TILING_TEMPLATE("Pow", PowTensorScalarTiling, 0);146+REGISTER_OPS_TILING_TEMPLATE(Pow, PowTensorScalarTiling, 0);
147} // namespace optiling147} // namespace optiling
Mmath/pow/op_host/arch35/pow_tensor_tensor_tiling_arch35.cpp+1-1
@@ -210,5 +210,5 @@ ge::graphStatus PowTensorTensorTiling::PostTiling()
210 return ge::GRAPH_SUCCESS;210 return ge::GRAPH_SUCCESS;
211}211}
212 212 
213-REGISTER_TILING_TEMPLATE("Pow", PowTensorTensorTiling, 1);213+REGISTER_OPS_TILING_TEMPLATE(Pow, PowTensorTensorTiling, 1);
214} // namespace optiling214} // namespace optiling
Mmath/reduce_mean/examples/test_aclnn_global_average_pool.cpp+5-5
@@ -101,14 +101,14 @@ int main() {
101 aclTensor* self = nullptr;101 aclTensor* self = nullptr;
102 aclTensor* out = nullptr;102 aclTensor* out = nullptr;
103 103 
104- std::vector<double> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7, 5, 3, 7, 6};104+ std::vector<float> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7, 5, 3, 7, 6};
105- std::vector<double> outHostData = {4.5, 7.25, 5.25};105+ std::vector<float> outHostData = {4.5, 7.25, 5.25};
106 106 
107 // 创建self aclTensor107 // 创建self aclTensor
108- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);108+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
109 CHECK_RET(ret == ACL_SUCCESS, return ret);109 CHECK_RET(ret == ACL_SUCCESS, return ret);
110 // 创建out aclTensor110 // 创建out aclTensor
111- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);111+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
112 CHECK_RET(ret == ACL_SUCCESS, return ret);112 CHECK_RET(ret == ACL_SUCCESS, return ret);
113 113 
114 // 3. 调用CANN算子库API114 // 3. 调用CANN算子库API
@@ -133,7 +133,7 @@ int main() {
133 133 
134 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改134 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
135 auto size = GetShapeSize(outShape);135 auto size = GetShapeSize(outShape);
136- std::vector<double> resultData(size, 0);136+ std::vector<float> resultData(size, 0);
137 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]),137 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]),
138 outDeviceAddr, size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);138 outDeviceAddr, size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
139 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy resultData from device to host failed. ERROR: %d\n", ret); return ret);139 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy resultData from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/reduce_mean/examples/test_aclnn_mean.cpp+5-5
@@ -81,18 +81,18 @@ int PrepareInputAndOutput(
81 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self, aclIntArray** dim,81 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self, aclIntArray** dim,
82 void** outDeviceAddr, aclTensor** out)82 void** outDeviceAddr, aclTensor** out)
83{83{
84- std::vector<int64_t> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7};84+ std::vector<float> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7};
85- std::vector<int64_t> outHostData = {2, 3, 5, 8};85+ std::vector<float> outHostData = {2, 3, 5, 8};
86 std::vector<int64_t> dimData = {1, 2};86 std::vector<int64_t> dimData = {1, 2};
87 87 
88 // 创建self aclTensor88 // 创建self aclTensor
89- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_INT64, self);89+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
90 CHECK_RET(ret == ACL_SUCCESS, return ret);90 CHECK_RET(ret == ACL_SUCCESS, return ret);
91 // 创建dim aclIntArray91 // 创建dim aclIntArray
92 ret = CreateAclIntArray(dimData, dim);92 ret = CreateAclIntArray(dimData, dim);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 // 创建out aclTensor94 // 创建out aclTensor
95- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_INT64, out);95+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
96 CHECK_RET(ret == ACL_SUCCESS, return ret);96 CHECK_RET(ret == ACL_SUCCESS, return ret);
97 97 
98 return ACL_SUCCESS;98 return ACL_SUCCESS;
@@ -145,7 +145,7 @@ int main() {
145 bool keepdim = false;145 bool keepdim = false;
146 aclOpExecutor* executor;146 aclOpExecutor* executor;
147 // 调用aclnnMean第一段接口147 // 调用aclnnMean第一段接口
148- ret = aclnnMeanGetWorkspaceSize(self, dim, keepdim, aclDataType::ACL_INT64, out, &workspaceSize, &executor);148+ ret = aclnnMeanGetWorkspaceSize(self, dim, keepdim, aclDataType::ACL_FLOAT, out, &workspaceSize, &executor);
149 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMeanGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);149 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnMeanGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
150 // 根据第一段接口计算出的workspaceSize申请device内存150 // 根据第一段接口计算出的workspaceSize申请device内存
151 void* workspaceAddr = nullptr;151 void* workspaceAddr = nullptr;
Mmath/reduce_mean/examples/test_aclnn_mean_v2.cpp+4-4
@@ -73,18 +73,18 @@ int PrepareInputAndOutput(
73 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self, aclIntArray** dim,73 std::vector<int64_t>& selfShape, std::vector<int64_t>& outShape, void** selfDeviceAddr, aclTensor** self, aclIntArray** dim,
74 void** outDeviceAddr, aclTensor** out)74 void** outDeviceAddr, aclTensor** out)
75{75{
76- std::vector<int64_t> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7};76+ std::vector<float> selfHostData = {2, 3, 5, 8, 4, 12, 6, 7};
77- std::vector<int64_t> outHostData = {2, 3, 5, 8};77+ std::vector<float> outHostData = {2, 3, 5, 8};
78 std::vector<int64_t> dimData = {1, 2};78 std::vector<int64_t> dimData = {1, 2};
79 79 
80 // 创建self aclTensor80 // 创建self aclTensor
81- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_INT64, self);81+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
82 CHECK_RET(ret == ACL_SUCCESS, return ret);82 CHECK_RET(ret == ACL_SUCCESS, return ret);
83 // 创建dim aclIntArray83 // 创建dim aclIntArray
84 *dim = aclCreateIntArray(dimData.data(), 1);84 *dim = aclCreateIntArray(dimData.data(), 1);
85 CHECK_RET(ret == ACL_SUCCESS, return false);85 CHECK_RET(ret == ACL_SUCCESS, return false);
86 // 创建out aclTensor86 // 创建out aclTensor
87- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_INT64, out);87+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
88 CHECK_RET(ret == ACL_SUCCESS, return ret);88 CHECK_RET(ret == ACL_SUCCESS, return ret);
89 89 
90 return ACL_SUCCESS;90 return ACL_SUCCESS;
Mmath/round/examples/test_aclnn_inplace_round.cpp+5-5
@@ -85,14 +85,14 @@ int main() {
85 aclTensor* self = nullptr;85 aclTensor* self = nullptr;
86 aclTensor* out = nullptr;86 aclTensor* out = nullptr;
87 87 
88- std::vector<double> selfHostData = {1.1, 2.8, 3.5};88+ std::vector<float> selfHostData = {1.1, 2.8, 3.5};
89- std::vector<double> outHostData = {0, 0, 0};89+ std::vector<float> outHostData = {0, 0, 0};
90 90 
91 // 创建self aclTensor91 // 创建self aclTensor
92- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);92+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 // 创建out aclTensor94 // 创建out aclTensor
95- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);95+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
96 CHECK_RET(ret == ACL_SUCCESS, return ret);96 CHECK_RET(ret == ACL_SUCCESS, return ret);
97 97 
98 uint64_t workspaceSize = 0;98 uint64_t workspaceSize = 0;
@@ -115,7 +115,7 @@ int main() {
115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
117 auto size = GetShapeSize(outShape);117 auto size = GetShapeSize(outShape);
118- std::vector<double> resultData(size, 0);118+ std::vector<float> resultData(size, 0);
119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr, 119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/round/examples/test_aclnn_inplace_round_decimals.cpp+5-5
@@ -82,15 +82,15 @@ int main() {
82 void* outDeviceAddr = nullptr;82 void* outDeviceAddr = nullptr;
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclTensor* out = nullptr;84 aclTensor* out = nullptr;
85- std::vector<double> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};85+ std::vector<float> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};
86- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};86+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
87 int decimals = 0;87 int decimals = 0;
88 88 
89 // 创建self aclTensor89 // 创建self aclTensor
90- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);90+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 // 创建out aclTensor92 // 创建out aclTensor
93- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);93+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
94 CHECK_RET(ret == ACL_SUCCESS, return ret);94 CHECK_RET(ret == ACL_SUCCESS, return ret);
95 95 
96 // 3. 调用CANN算子库API，需要修改为具体的Api名称96 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -115,7 +115,7 @@ int main() {
115 115 
116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
117 auto size = GetShapeSize(outShape);117 auto size = GetShapeSize(outShape);
118- std::vector<double> resultData(size, 0);118+ std::vector<float> resultData(size, 0);
119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/round/examples/test_aclnn_round.cpp+5-5
@@ -85,14 +85,14 @@ int main() {
85 aclTensor* self = nullptr;85 aclTensor* self = nullptr;
86 aclTensor* out = nullptr;86 aclTensor* out = nullptr;
87 87 
88- std::vector<double> selfHostData = {1.1, 2.8, 3.5};88+ std::vector<float> selfHostData = {1.1, 2.8, 3.5};
89- std::vector<double> outHostData = {0, 0, 0};89+ std::vector<float> outHostData = {0, 0, 0};
90 90 
91 // 创建self aclTensor91 // 创建self aclTensor
92- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);92+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 // 创建out aclTensor94 // 创建out aclTensor
95- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);95+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
96 CHECK_RET(ret == ACL_SUCCESS, return ret);96 CHECK_RET(ret == ACL_SUCCESS, return ret);
97 97 
98 uint64_t workspaceSize = 0;98 uint64_t workspaceSize = 0;
@@ -115,7 +115,7 @@ int main() {
115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);115 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);
116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
117 auto size = GetShapeSize(outShape);117 auto size = GetShapeSize(outShape);
118- std::vector<double> resultData(size, 0);118+ std::vector<float> resultData(size, 0);
119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, 119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/round/examples/test_aclnn_round_decimals.cpp+5-5
@@ -82,15 +82,15 @@ int main() {
82 void* outDeviceAddr = nullptr;82 void* outDeviceAddr = nullptr;
83 aclTensor* self = nullptr;83 aclTensor* self = nullptr;
84 aclTensor* out = nullptr;84 aclTensor* out = nullptr;
85- std::vector<double> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};85+ std::vector<float> selfHostData = {0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7,0.8};
86- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};86+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
87 int decimals = 0;87 int decimals = 0;
88 88 
89 // 创建self aclTensor89 // 创建self aclTensor
90- ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_DOUBLE, &self);90+ ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
91 CHECK_RET(ret == ACL_SUCCESS, return ret);91 CHECK_RET(ret == ACL_SUCCESS, return ret);
92 // 创建out aclTensor92 // 创建out aclTensor
93- ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_DOUBLE, &out);93+ ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
94 CHECK_RET(ret == ACL_SUCCESS, return ret);94 CHECK_RET(ret == ACL_SUCCESS, return ret);
95 95 
96 // 3. 调用CANN算子库API，需要修改为具体的Api名称96 // 3. 调用CANN算子库API，需要修改为具体的Api名称
@@ -115,7 +115,7 @@ int main() {
115 115 
116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改116 // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧，需要根据具体API的接口定义修改
117 auto size = GetShapeSize(outShape);117 auto size = GetShapeSize(outShape);
118- std::vector<double> resultData(size, 0);118+ std::vector<float> resultData(size, 0);
119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,119 ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);120 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);121 CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
Mmath/sinh/examples/test_aclnn_sinh.cpp+7-7
@@ -104,7 +104,7 @@ int ExecuteInplaceSinh(aclTensor* self, aclrtStream stream) {
104int PrintResults(void* outDeviceAddr, std::vector<int64_t>& outShape,104int PrintResults(void* outDeviceAddr, std::vector<int64_t>& outShape,
105 void* selfDeviceAddr, std::vector<int64_t>& selfShape) {105 void* selfDeviceAddr, std::vector<int64_t>& selfShape) {
106 auto size = GetShapeSize(outShape);106 auto size = GetShapeSize(outShape);
107- std::vector<double> resultData(size, 0);107+ std::vector<float> resultData(size, 0);
108 int ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,108 int ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
109 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);109 size * sizeof(resultData[0]), ACL_MEMCPY_DEVICE_TO_HOST);
110 if (ret != ACL_SUCCESS) return ret;110 if (ret != ACL_SUCCESS) return ret;
@@ -114,9 +114,9 @@ int PrintResults(void* outDeviceAddr, std::vector<int64_t>& outShape,
114 }114 }
115 115 
116 auto inplaceSize = GetShapeSize(selfShape);116 auto inplaceSize = GetShapeSize(selfShape);
117- std::vector<double> inplaceResultData(inplaceSize, 0);117+ std::vector<float> inplaceResultData(inplaceSize, 0);
118 ret = aclrtMemcpy(inplaceResultData.data(), inplaceResultData.size() * sizeof(inplaceResultData[0]),118 ret = aclrtMemcpy(inplaceResultData.data(), inplaceResultData.size() * sizeof(inplaceResultData[0]),
119- selfDeviceAddr, inplaceSize * sizeof(double), ACL_MEMCPY_DEVICE_TO_HOST);119+ selfDeviceAddr, inplaceSize * sizeof(float), ACL_MEMCPY_DEVICE_TO_HOST);
120 if (ret != ACL_SUCCESS) return ret;120 if (ret != ACL_SUCCESS) return ret;
121 121 
122 for (int64_t i = 0; i < inplaceSize; i++) {122 for (int64_t i = 0; i < inplaceSize; i++) {
@@ -130,14 +130,14 @@ int CreateInputOutputTensors(aclTensor** self, aclTensor** out,
130 void** selfDeviceAddr, void** outDeviceAddr) {130 void** selfDeviceAddr, void** outDeviceAddr) {
131 std::vector<int64_t> selfShape = {4, 2};131 std::vector<int64_t> selfShape = {4, 2};
132 std::vector<int64_t> outShape = {4, 2};132 std::vector<int64_t> outShape = {4, 2};
133- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};133+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
134- std::vector<double> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};134+ std::vector<float> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
135 135 
136 // 创建self aclTensor136 // 创建self aclTensor
137- int ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);137+ int ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
138 CHECK_RET(ret == ACL_SUCCESS, return ret);138 CHECK_RET(ret == ACL_SUCCESS, return ret);
139 // 创建out aclTensor139 // 创建out aclTensor
140- ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_DOUBLE, out);140+ ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_FLOAT, out);
141 CHECK_RET(ret == ACL_SUCCESS, return ret);141 CHECK_RET(ret == ACL_SUCCESS, return ret);
142 142 
143 return ACL_SUCCESS;143 return ACL_SUCCESS;
Mmath/tensor_equal/examples/test_aclnn_equal.cpp+4-4
@@ -85,14 +85,14 @@ aclError CreateInputs(
85 void** selfDeviceAddr, void** otherDeviceAddr, void** outDeviceAddr, aclTensor** self, aclTensor** other,85 void** selfDeviceAddr, void** otherDeviceAddr, void** outDeviceAddr, aclTensor** self, aclTensor** other,
86 aclTensor** out)86 aclTensor** out)
87{87{
88- std::vector<double> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};88+ std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
89- std::vector<double> otherHostData = {0, 1, 2, 3, 4, 5, 6, 7};89+ std::vector<float> otherHostData = {0, 1, 2, 3, 4, 5, 6, 7};
90 std::vector<char> outHostData = {0};90 std::vector<char> outHostData = {0};
91 91 
92- auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_DOUBLE, self);92+ auto ret = CreateAclTensor(selfHostData, selfShape, selfDeviceAddr, aclDataType::ACL_FLOAT, self);
93 CHECK_RET(ret == ACL_SUCCESS, return ret);93 CHECK_RET(ret == ACL_SUCCESS, return ret);
94 94 
95- ret = CreateAclTensor(otherHostData, otherShape, otherDeviceAddr, aclDataType::ACL_DOUBLE, other);95+ ret = CreateAclTensor(otherHostData, otherShape, otherDeviceAddr, aclDataType::ACL_FLOAT, other);
96 CHECK_RET(ret == ACL_SUCCESS, return ret);96 CHECK_RET(ret == ACL_SUCCESS, return ret);
97 97 
98 ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_BOOL, out);98 ret = CreateAclTensor(outHostData, outShape, outDeviceAddr, aclDataType::ACL_BOOL, out);