aclnnForeachAddcmulScalarV2

Applicable Products

Product	Supported
Ascend 950PR/Ascend 950DT	√
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	×
Atlas inference products	×
Atlas training products	×

Function

Description: Performs element-wise multiplication on tensor list x2 and tensor list x3, then multiplies by the scalar, and finally performs element-wise addition of the previously computed result with tensor list x1. Compared with aclnnForeachAddcmulScalar, this API changes the structure type of the input parameter scalar from aclTensor to aclScalar. Select the appropriate API based on your actual situation.
Formula:
$x1 = [{x1_0}, {x1_1}, ... {x1_{n-1}}], x2 = [{x2_0}, {x2_1}, ... {x2_{n-1}}], x3 = [{x3_0}, {x3_1}, ... {x3_{n-1}}]\\ y = [{y_0}, {y_1}, ... {y_{n-1}}]\\$ ${\rm y}_i = x1_{i} + {\rm scalar} × x2_{i} × x3_{i} (i=0,1,...n-1)$

Function Prototype

Each operator is divided into a two-phase API. You must first call the "aclnnForeachAddcmulScalarV2GetWorkspaceSize" API to obtain the input parameters and calculate the required workspace size based on the computation process, and then call the "aclnnForeachAddcmulScalarV2" API to perform the computation.

aclnnStatus aclnnForeachAddcmulScalarV2GetWorkspaceSize(
  const aclTensorList *x1,
  const aclTensorList *x2,
  const aclTensorList *x3,
  const aclScalar     *scalar,
  aclTensorList       *out,
  uint64_t            *workspaceSize,
  aclOpExecutor      **executor)

aclnnStatus aclnnForeachAddcmulScalarV2(
  void          *workspace,
  uint64_t       workspaceSize,
  aclOpExecutor *executor,
  aclrtStream    stream)

aclnnForeachAddcmulScalarV2GetWorkspaceSize

Parameters

Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (Shape)	Non-Contiguous Tensor
x1 (aclTensorList*)	Input	Represents the first input tensor list for addition in the mixed operation. Corresponds to `x1` in the formula.	Empty tensors are not supported. All tensors in this parameter must have the same data type.	FLOAT32, FLOAT16, BFLOAT16, INT32	ND	0-8	×
x2 (aclTensorList*)	Input	Represents the second input tensor list for the multiplication in the mixed operation. Corresponds to `x2` in the formula.	Empty tensors are not supported. All tensors in this parameter must have the same data type. The data type, data format, and shape must be consistent with the input parameter `x1`.	FLOAT32, FLOAT16, BFLOAT16, INT32	ND	0-8	×
x3 (aclTensorList*)	Input	Represents the third input tensor list for the multiplication in the mixed operation. Corresponds to `x3` in the formula.	Empty tensors are not supported. All tensors in this parameter must have the same data type. The data type, data format, and shape must be consistent with the input parameter `x1`.	FLOAT32, FLOAT16, BFLOAT16, INT32	ND	0-8	×
scalar (aclScalar*)	Input	Represents the first input scalar for the multiplication in the mixed operation. Corresponds to `scalar` in the formula.	-	FLOAT32, FLOAT16, INT32, DOUBLE, INT64	-	-	-
out (aclTensorList*)	Output	Represents the output tensor list of the mixed operation. Corresponds to `y` in the formula.	Empty tensors are not supported. All tensors in this parameter must have the same data type. The data type and data format must be consistent with those of the input `x1`, and the shape size must be greater than or equal to the shape size of the input `x1`.	FLOAT32, FLOAT16, BFLOAT16, INT32	ND	0-8	×
workspaceSize (uint64_t*)	Output	Returns the workspace size to be requested on the Device side.	-	-	-	-	-
executor (aclOpExecutor**)	Output	Returns the op executor, which contains the operator computation flow.	-	-	-	-	-

Atlas A2 training products/Atlas A2 inference products, Atlas A3 training products/Atlas A3 inference products:

The data type of scalar has a certain correspondence with the data type of the input x1:
- When the data type of x1 is FLOAT32 or BFLOAT16, the data type of scalar supports FLOAT32 and DOUBLE.
- When the data type of x1 is FLOAT16, the data type of scalar supports FLOAT16 and DOUBLE.
- When the data type of x1 is INT32, the data type of scalar supports INT32 and INT64.
Ascend 950PR/Ascend 950DT:
- The data type of scalar has a specific correspondence with the data type of the input x1:
  - When the data type of x1 is BFLOAT16 or FLOAT32, the data type of scalar supports FLOAT32 and DOUBLE.
  - When the data type of x1 is FLOAT16, the data type of scalar supports FLOAT16, FLOAT32, and DOUBLE.
  - When the data type of x1 is INT32, the data type of scalar supports INT32 and INT64.
- The maximum number of tensors supported by the input x1, x2, x3, and the output y is 50 each.

Return Value

aclnnStatus: return code. For details, see aclnn Return Code.

The first-phase API performs input parameter validation and returns an error in the following scenarios:

Return Code	Error Code	Description
ACLNN_ERR_PARAM_NULLPTR	161001	The input x1, x2, x3, scalar, and out are null pointers.
ACLNN_ERR_PARAM_INVALID	161002	The data types of x1, x2, x3, scalar, and out are not within the supported range.
		The data types of x1, x2, x3, and out are inconsistent.
		The shapes of x1, x2, x3, and out do not satisfy the constraints.
		The element data types of tensors in x1, x2, x3, or out are inconsistent.
		A tensor in x1, x2, x3, or out has more than 8 dimensions.

aclnnForeachAddcmulScalarV2

Parameters

Parameter	Input/Output	Description
workspace	Input	Specifies the workspace memory address applied for on the device side.
workspaceSize	Input	Specifies the workspace size applied for on the device side, obtained through the first-phase API aclnnForeachAddcmulScalarV2GetWorkspaceSize.
executor	Input	Operator executor that contains the operator computation flow.
stream	Input	Specifies the stream for task execution.

Return Value

aclnnStatus: return code. For details, see aclnn Return Code.

Constraints

Deterministic computation:
- aclnnForeachAddcmulScalarV2 defaults to a deterministic implementation.

Example

The following provides example code for reference only. For details about compilation and running, see Compile and Run Samples.

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_foreach_addcmul_scalar_v2.h"

#define CHECK_RET(cond, return_expr) \
  do {                               \
    if (!(cond)) {                   \
      return_expr;                   \
    }                                \
  } while (0)

#define LOG_PRINT(message, ...)     \
  do {                              \
    printf(message, ##__VA_ARGS__); \
  } while (0)

int64_t GetShapeSize(const std::vector<int64_t>& shape) {
  int64_t shapeSize = 1;
  for (auto i : shape) {
    shapeSize *= i;
  }
  return shapeSize;
}

int Init(int32_t deviceId, aclrtStream *stream)
{
    // Boilerplate for resource initialization.
    auto ret = aclInit(nullptr);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret);
    ret = aclrtSetDevice(deviceId);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret);
    ret = aclrtCreateStream(stream);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret);
    return 0;
}

template <typename T>
int CreateAclTensor(const std::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr,
                    aclDataType dataType, aclTensor** tensor) {
  auto size = GetShapeSize(shape) * sizeof(T);
  // Call aclrtMalloc to apply for device-side memory.
  auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret);
  // Copy host-side data to device memory using aclrtMemcpy.
  ret = aclrtMemcpy(*deviceAddr, size, hostData.data(), size, ACL_MEMCPY_HOST_TO_DEVICE);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMemcpy failed. ERROR: %d\n", ret); return ret);

  // Compute the strides of a contiguous tensor.
  std::vector<int64_t> strides(shape.size(), 1);
  for (int64_t i = shape.size() - 2; i >= 0; i--) {
    strides[i] = shape[i + 1] * strides[i + 1];
  }

  // Call the aclCreateTensor API to create an aclTensor.
  *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
                            shape.data(), shape.size(), *deviceAddr);
  return 0;
}

int main() {
  // 1. (Boilerplate) Initialize the device/stream. See the ACL API manual.
  // Set deviceId based on the actual device.
  int32_t deviceId = 0;
  aclrtStream stream;
  auto ret = Init(deviceId, &stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);

  // 2. Construct the inputs and outputs. Customize the construction based on the API interface.
  std::vector<int64_t> selfShape1 = {2, 3};
  std::vector<int64_t> selfShape2 = {1, 3};
  std::vector<int64_t> otherShape1 = {2, 3};
  std::vector<int64_t> otherShape2 = {1, 3};
  std::vector<int64_t> anotherShape1 = {2, 3};
  std::vector<int64_t> anotherShape2 = {1, 3};
  std::vector<int64_t> outShape1 = {2, 3};
  std::vector<int64_t> outShape2 = {1, 3};
  void* input1DeviceAddr = nullptr;
  void* input2DeviceAddr = nullptr;
  void* other1DeviceAddr = nullptr;
  void* other2DeviceAddr = nullptr;
  void* another1DeviceAddr = nullptr;
  void* another2DeviceAddr = nullptr;
  void* out1DeviceAddr = nullptr;
  void* out2DeviceAddr = nullptr;
  aclTensor* input1 = nullptr;
  aclTensor* input2 = nullptr;
  aclTensor* other1 = nullptr;
  aclTensor* other2 = nullptr;
  aclTensor* another1 = nullptr;
  aclTensor* another2 = nullptr;
  aclScalar* alpha = nullptr;
  aclTensor* out1 = nullptr;
  aclTensor* out2 = nullptr;
  std::vector<float> input1HostData = {1, 2, 3, 4, 5, 6};
  std::vector<float> input2HostData = {7, 8, 9};
  std::vector<float> other1HostData = {1, 2, 3, 4, 5, 6};
  std::vector<float> other2HostData = {7, 8, 9};
  std::vector<float> another1HostData = {1, 2, 3, 4, 5, 6};
  std::vector<float> another2HostData = {7, 8, 9};
  std::vector<float> out1HostData(6, 0);
  std::vector<float> out2HostData(3, 0);
  float alphaValue = 1.2f;
  // Create the input1 aclTensor.
  ret = CreateAclTensor(input1HostData, selfShape1, &input1DeviceAddr, aclDataType::ACL_FLOAT, &input1);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the input2 aclTensor.
  ret = CreateAclTensor(input2HostData, selfShape2, &input2DeviceAddr, aclDataType::ACL_FLOAT, &input2);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the other1 aclTensor.
  ret = CreateAclTensor(other1HostData, otherShape1, &other1DeviceAddr, aclDataType::ACL_FLOAT, &other1);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create the other2 aclTensor.
  ret = CreateAclTensor(other2HostData, otherShape2, &other2DeviceAddr, aclDataType::ACL_FLOAT, &other2);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create another1 aclTensor.
  ret = CreateAclTensor(another1HostData, anotherShape1, &another1DeviceAddr, aclDataType::ACL_FLOAT, &another1);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create another2 aclTensor.
  ret = CreateAclTensor(another2HostData, anotherShape2, &another2DeviceAddr, aclDataType::ACL_FLOAT, &another2);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create alpha aclScalar.
  alpha = aclCreateScalar(&alphaValue, aclDataType::ACL_FLOAT);
  CHECK_RET(alpha != nullptr, return ret);
  // Create out1 aclTensor.
  ret = CreateAclTensor(out1HostData, outShape1, &out1DeviceAddr, aclDataType::ACL_FLOAT, &out1);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // Create out2 aclTensor.
  ret = CreateAclTensor(out2HostData, outShape2, &out2DeviceAddr, aclDataType::ACL_FLOAT, &out2);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  std::vector<aclTensor*> tempInput1{input1, input2};
  aclTensorList* tensorListInput1 = aclCreateTensorList(tempInput1.data(), tempInput1.size());
  std::vector<aclTensor*> tempInput2{other1, other2};
  aclTensorList* tensorListInput2 = aclCreateTensorList(tempInput2.data(), tempInput2.size());
  std::vector<aclTensor*> tempAnother{another1, another2};
  aclTensorList* tensorListAnother = aclCreateTensorList(tempAnother.data(), tempAnother.size());
  std::vector<aclTensor*> tempOutput{out1, out2};
  aclTensorList* tensorListOutput = aclCreateTensorList(tempOutput.data(), tempOutput.size());

  // 3. Call the CANN operator library API. Replace with the actual API name.
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // Call the first-phase API of aclnnForeachAddcmulScalarV2.
  ret = aclnnForeachAddcmulScalarV2GetWorkspaceSize(tensorListInput1, tensorListInput2, tensorListAnother, alpha, tensorListOutput, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnForeachAddcmulScalarV2GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // Apply for device memory based on the workspaceSize calculated by the first-phase API.
  void* workspaceAddr = nullptr;
  if (workspaceSize > 0) {
    ret = aclrtMalloc(&workspaceAddr, workspaceSize, ACL_MEM_MALLOC_HUGE_FIRST);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("allocate workspace failed. ERROR: %d\n", ret); return ret);
  }
  // Call the second-phase API of aclnnForeachAddcmulScalarV2.
  ret = aclnnForeachAddcmulScalarV2(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnForeachAddcmulScalarV2 failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate) Synchronize and wait for task execution to complete.
  ret = aclrtSynchronizeStream(stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);

  // 5. Obtain the output value and copy the result from the device memory to the host memory. Modify based on the specific API definition.
  auto size = GetShapeSize(outShape1);
  std::vector<float> out1Data(size, 0);
  ret = aclrtMemcpy(out1Data.data(), out1Data.size() * sizeof(out1Data[0]), out1DeviceAddr,
                    size * sizeof(out1Data[0]), ACL_MEMCPY_DEVICE_TO_HOST);
  for (int64_t i = 0; i < size; i++) {
    LOG_PRINT("out1 result[%ld] is: %f\n", i, out1Data[i]);
  }

  size = GetShapeSize(outShape2);
  std::vector<float> out2Data(size, 0);
  ret = aclrtMemcpy(out2Data.data(), out2Data.size() * sizeof(out2Data[0]), out2DeviceAddr,
                    size * sizeof(out2Data[0]), ACL_MEMCPY_DEVICE_TO_HOST);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("copy result from device to host failed. ERROR: %d\n", ret); return ret);
  for (int64_t i = 0; i < size; i++) {
    LOG_PRINT("out2 result[%ld] is: %f\n", i, out2Data[i]);
  }

  // 6. Release aclTensor and aclScalar. Modify based on the specific API definition.
  aclDestroyTensorList(tensorListInput1);
  aclDestroyTensorList(tensorListInput2);
  aclDestroyTensorList(tensorListAnother);
  aclDestroyTensorList(tensorListOutput);
  aclDestroyScalar(alpha);

  // 7. Release device resources. Modify based on the specific API definition.
  aclrtFree(input1DeviceAddr);
  aclrtFree(input2DeviceAddr);
  aclrtFree(other1DeviceAddr);
  aclrtFree(other2DeviceAddr);
  aclrtFree(another1DeviceAddr);
  aclrtFree(another2DeviceAddr);
  aclrtFree(out1DeviceAddr);
  aclrtFree(out2DeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
  return 0;
}