aclnnEqScalar&aclnnInplaceEqScalar

支持的产品型号

• Atlas 训练系列产品。
• Atlas A2训练系列产品/Atlas 800I A2推理产品。

接口原型

• aclnnEqScalar和aclnnInplaceEqScalar实现相同的功能，使用区别如下，请根据自身实际场景选择合适的算子。

  • aclnnEqScalar：需新建一个输出张量对象存储计算结果。
  • aclnnInplaceEqScalar：无需新建输出张量对象，直接在输入张量的内存中存储计算结果。

• 每个算子分为两段式接口，必须先调用“aclnnEqScalarGetWorkspaceSize”或者“aclnnInplaceEqScalarGetWorkspaceSize”接口获取计算所需workspace大小以及包含了算子计算流程的执行器，再调用“aclnnEqScalar”或者“aclnnInplaceEqScalar”接口执行计算。

  • aclnnStatus aclnnEqScalarGetWorkspaceSize(const aclTensor *self, const aclScalar *other, aclTensor *out, uint64_t *workspaceSize, aclOpExecutor **executor)
  • aclnnStatus aclnnEqScalar(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream)
  • aclnnStatus aclnnInplaceEqScalarGetWorkspaceSize(const aclTensor *selfRef, const aclScalar *other, uint64_t *workspaceSize, aclOpExecutor **executor)
  • aclnnStatus aclnnInplaceEqScalar(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream)

功能描述

• 算子功能：计算self中的元素的值与other的值是否相等，将self每个元素与other的值的比较结果写入out中。
• 计算公式：$$out_i = (self_i == \mathit{other} ) ? [True] : [False]$$

aclnnEqScalarGetWorkspaceSize

• 参数说明：

  • self(const aclTensor *, 计算输入)：Device侧的aclTensor，数据类型支持FLOAT16, FLOAT, INT64, UINT64(仅Atlas 训练系列产品支持), INT32, INT8, UINT8, BOOL, UINT32(仅Atlas 训练系列产品支持), BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）, DOUBLE, INT16, COMPLEX64, COMPLEX128数据类型，数据格式支持ND。
  • other(const aclScalar *, 计算输入)：Host侧的aclScalar，数据类型支持FLOAT16, FLOAT, INT64, UINT64(仅Atlas 训练系列产品支持), INT32, INT8, UINT8, BOOL, UINT32(仅Atlas 训练系列产品支持), BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）, DOUBLE, INT16, COMPLEX64, COMPLEX128数据类型。
  • out(aclTensor *, 计算输出)：Device侧的aclTensor，数据类型为bool, 或者Cast能支持转换的数据类型, shape与self的shape一致，数据格式支持ND。
  • workspaceSize(uint64_t *, 出参)：返回需要在Device侧申请的workspace大小。
  • executor(aclOpExecutor **, 出参)：返回op执行器，包含了算子计算流程。

• 返回值：

  aclnnStatus：返回状态码，具体参见aclnn返回码。

  返回161001（ACLNN_ERR_PARAM_NULLPTR）：1. 传入的self、other、out是空指针时。
  返回161002（ACLNN_ERR_PARAM_INVALID）：1. self和other的数据类型不在支持的范围之内。
                                        2. self和out的shape不一致。

aclnnEqScalar

• 参数说明：

  • workspace(void *, 入参)：在Device侧申请的workspace内存地址。
  • workspaceSize(uint64_t, 入参)：在Device侧申请的workspace大小，由第一段接口aclnnEqScalarGetWorkspaceSize获取。
  • executor(aclOpExecutor *, 入参)：op执行器，包含了算子计算流程。
  • stream(aclrtStream, 入参)：指定执行任务的 AscendCL Stream流。

• 返回值：

  aclnnStatus：返回状态码，具体参见aclnn返回码。

aclnnInplaceEqScalarGetWorkspaceSize

• 参数说明：

  • selfRef(const aclTensor *，计算输入|计算输出)：输入输出tensor，即公式中的self与out，Device侧的aclTensor，数据类型支持FLOAT16, FLOAT, INT64, UINT64(仅Atlas 训练系列产品支持), INT32, INT8, UINT8, BOOL, UINT32(仅Atlas 训练系列产品支持), BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）, DOUBLE, INT16, COMPLEX64, COMPLEX128数据类型，数据格式支持ND。
  • other(const aclScalar *，计算输入)：Host侧的aclScalar，数据类型支持FLOAT16, FLOAT, INT64, UINT64(仅Atlas 训练系列产品支持), INT32, INT8, UINT8, BOOL, UINT32(仅Atlas 训练系列产品支持), BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）, DOUBLE, INT16, COMPLEX64, COMPLEX128数据类型，且数据类型与selfRef的数据类型需满足数据类型推导规则（参见互推导关系）。
  • workspaceSize(uint64_t *，出参)：返回需要在Device侧申请的workspace大小。
  • executor(aclOpExecutor **，出参)：返回op执行器，包含了算子计算流程。

• 返回值：

  aclnnStatus：返回状态码，具体参见aclnn返回码。

  返回161001（ACLNN_ERR_PARAM_NULLPTR）：1. 传入的selfRef和other是空指针时。
  返回161002（ACLNN_ERR_PARAM_INVALID）：1. selfRef和other的数据类型不在支持的范围之内。

aclnnInplaceEqScalar

• 参数说明：

  • workspace(void *, 入参)：在Device侧申请的workspace内存地址。
  • workspaceSize(uint64_t, 入参)：在Device侧申请的workspace大小，由第一段接口aclnnInplaceEqScalarGetWorkspaceSize获取。
  • executor(aclOpExecutor *, 入参)：op执行器，包含了算子计算流程。
  • stream(aclrtStream, 入参)：指定执行任务的 AscendCL Stream流。

• 返回值：

  aclnnStatus：返回状态码，具体参见aclnn返回码。

约束与限制

无

调用示例

示例代码如下，仅供参考，具体编译和执行过程请参考编译与运行样例。

aclnnEqScalar示例代码：

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_eq_scalar.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 shape_size = 1;
  for (auto i : shape) {
    shape_size *= i;
  }
  return shape_size;
}

int Init(int32_t deviceId, aclrtStream* stream) {
  // 固定写法，AscendCL初始化
  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);
  // 调用aclrtMalloc申请device侧内存
  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);

  // 调用aclrtMemcpy将host侧数据拷贝到device侧内存上
  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);

  // 计算连续tensor的strides
  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];
  }

  // 调用aclCreateTensor接口创建aclTensor
  *tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
                            shape.data(), shape.size(), *deviceAddr);
  return 0;
}

int main() {
  int32_t deviceId = 0;
  aclrtStream stream;
  auto ret = Init(deviceId, &stream);
  CHECK_RET(ret == 0, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);
  std::vector<int64_t> selfShape = {4, 2};
  std::vector<int64_t> outShape = {4, 2};

  void* selfDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclScalar* other = nullptr;
  aclTensor* out = nullptr;
  std::vector<float> selfHostData = {0, 1, 1.2, 0.3, 4.1, 5, 1.6, 7};
  std::vector<char> outHostData = {0, 0, 0, 0, 0, 0, 0, 0};
  float otherValue = 1.2f;

  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
  CHECK_RET(other != nullptr, return ret);
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_BOOL, &out);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  ret = aclnnEqScalarGetWorkspaceSize(self, other, out, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnEqScalarGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);

  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);
  }

  ret = aclnnEqScalar(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnEqScalar failed. ERROR: %d\n", ret); return ret);

  ret = aclrtSynchronizeStream(stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);

  auto size = GetShapeSize(outShape);
  std::vector<char> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr, size * sizeof(char),
                    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("result[%ld] is: %d\n", i, resultData[i]);
  }

  aclDestroyTensor(self);
  aclDestroyScalar(other);
  aclDestroyTensor(out);

  aclrtFree(selfDeviceAddr);
  aclrtFree(outDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();

  return 0;
}

aclnnInplaceEqScalar示例代码：

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_eq_scalar.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) {
  // 固定写法，AscendCL初始化
  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);
  // 调用aclrtMalloc申请device侧内存
  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);
  // 调用aclrtMemcpy将host侧数据拷贝到device侧内存上
  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);

  // 计算连续tensor的strides
  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];
  }

  // 调用aclCreateTensor接口创建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. （固定写法）device/stream初始化，参考AscendCL对外接口列表
  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. 构造输入与输出，需要根据API的接口自定义构造
  std::vector<int64_t> selfShape = {4, 2};
  void* selfDeviceAddr = nullptr;
  aclTensor* self = nullptr;
  aclScalar* other = nullptr;
  std::vector<float> selfHostData = {0, 1, 2, 3, 4, 5, 6, 7};
  float otherValue = 2.0f;

  // 创建self aclTensor
  ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建other aclScalar
  other = aclCreateScalar(&otherValue, aclDataType::ACL_FLOAT);
  CHECK_RET(other != nullptr, return ret);

  // 3. 调用CANN算子库API
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // 调用aclnnInplaceEqScalar第一段接口
  ret = aclnnInplaceEqScalarGetWorkspaceSize(self, other, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnInplaceEqScalarGetWorkspaceSize failed. ERROR: %d\n", ret);
            return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  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);
  }
  // 调用aclnnInplaceEqScalar第二段接口
  ret = aclnnInplaceEqScalar(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnInplaceEqScalar failed. ERROR: %d\n", ret); return ret);

  // 4. （固定写法）同步等待任务执行结束
  ret = aclrtSynchronizeStream(stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);

  // 5. 获取输出的值，将device侧内存上的结果拷贝至host侧
  auto size = GetShapeSize(selfShape);
  std::vector<float> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), selfDeviceAddr,
                    size * sizeof(resultData[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("result[%ld] is: %f\n", i, resultData[i]);
  }

  // 6. 释放aclTensor和aclScalar
  aclDestroyTensor(self);
  aclDestroyScalar(other);

  // 7. 释放device资源
  aclrtFree(selfDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
  return 0;
}