昇腾社区首页
中文
注册
开发者
下载
支持

aclnnLayerNorm/aclnnLayerNormWithImplMode

接口原型

  • aclnnLayerNorm和aclnnLayerNormWithImplMode实现相同的功能，其使用区别在于后者提供了“implMode”计算模式选项，请根据自身实际场景选择合适的算子。
  • 每个算子分为两段接口，必须先调用“aclnnXxxGetWorkspaceSize”接口获取入参并根据计算流程计算所需workspace大小，再调用“aclnnXxx”接口执行计算。
  • aclnnLayerNorm两段式接口如下：
    • 第一段接口：aclnnStatus aclnnLayerNormGetWorkspaceSize(const aclTensor *input, const aclIntArray *normalizedShape, const aclTensor *weightOptional, const aclTensor *biasOptional, double eps, aclTensor *out, aclTensor *meanOutOptional, aclTensor *rstdOutOptional, uint64_t *workspaceSize, aclOpExecutor **executor)
    • 第二段接口：aclnnStatus aclnnLayerNorm(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream)
  • aclnnLayerNormWithImplMode两段式接口如下：
    • 第一段接口：aclnnStatus aclnnLayerNormWithImplModeGetWorkspaceSize(const aclTensor *input, const aclIntArray *normalizedShape, const aclTensor *weightOptional, const aclTensor *biasOptional, double eps, aclTensor *out, aclTensor *meanOutOptional, aclTensor *rstdOutOptional, int32_t implMode, uint64_t *workspaceSize, aclOpExecutor **executor)
    • 第二段接口：aclnnStatus aclnnLayerNormWithImplMode(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream)

功能描述

  • 算子功能：对指定层进行均值为0、标准差为1的归一化计算。
  • 计算公式：

    其中E[x]表示输入的均值，Var[x]表示输入的方差。

aclnnLayerNormGetWorkspaceSize

  • 接口定义：

    aclnnStatus aclnnLayerNormGetWorkspaceSize(const aclTensor *input, const aclIntArray *normalizedShape, const aclTensor *weightOptional, const aclTensor *biasOptional, double eps, aclTensor *out, aclTensor *meanOutOptional, aclTensor *rstdOutOptional, uint64_t *workspaceSize, aclOpExecutor **executor)

  • 参数说明：
    • input：Device侧的aclTensor，公式中的x，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape长度≥normalizedShape长度，且与normalizedShape右对齐时对应的维度shape相等。支持非连续的Tensor，数据格式支持ND。
    • normalizedShape：Host侧的aclIntArray，表示需要进行norm计算的shape，数据类型支持INT64。其长度小于等于input的shape长度，与input的shape右对齐时的维度shape相等。
    • weightOptional：Device侧的aclTensor，公式中的w，可选参数。weightOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与normalizedShape相等。支持非连续的Tensor，数据格式支持ND。weightOptional为空时，需要构造一个shape与normalizedShape相等、数据全为1的张量。
    • biasOptional：Device侧的aclTensor，公式中的bias，可选参数。biasOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与normalizedShape相等。支持非连续的Tensor，数据格式支持ND。biasOptional为空时，需要构造一个shape与normalizedShape相等、数据全为0的张量。
    • eps：Host侧的浮点数，公式中的eps，用于规避除零计算，数据类型为DOUBLE，需要是可转换成与input相同的数据类型。
    • out：Device侧的aclTensor，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与input的shape相等。支持非连续的Tensor，数据格式支持ND。
    • meanOutOptional：Device侧的aclTensor，可选参数。meanOutOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与rstdOutOptional的shape相等。支持非连续的Tensor，数据格式支持ND。
    • rstdOutOptional：Device侧的aclTensor，可选参数。rstdOutOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与meanOutOptional的shape相等。支持非连续的Tensor，数据格式支持ND。
    • workspaceSize：返回用户需要在Device侧申请的workspace大小。
    • executor：返回op执行器，包含了算子计算流程。
  • 返回值：

    返回aclnnStatus状态码，具体参见aclnn返回码。

    第一段接口完成入参校验，出现以下场景时报错：

    • 返回161001（ACLNN_ERR_PARAM_NULLPTR）：传入的input、normalizedShape或out为空指针。
    • 返回161002（ACLNN_ERR_PARAM_INVALID）：
      • input、normalizedShape、weightOptional非空时、biasOptional非空时、out、meanOutOptional非空时或rstdOutOptional非空时的shape超过8维。
      • input、weightOptional非空时、biasOptional非空时、out、meanOutOptional非空时或rstdOutOptional非空时的数据类型不在支持的范围内。
      • normalizedShape维度小于1维。
      • weightOptional非空且shape与normalizedShape不相等。
      • biasOptional非空且shape与normalizedShape不相等。
      • input的维度小于normalizedShape的维度。
      • input的shape与normalizedShape右对齐时对应维度shape不相等。

aclnnLayerNorm

  • 接口定义：

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

  • 参数说明：
    • workspace：在Device侧申请的workspace内存起址。
    • workspaceSize：在Device侧申请的workspace大小，由第一段接口aclnnLayerNormGetWorkspaceSize获取。
    • executor：op执行器，包含了算子计算流程。
    • stream：指定执行任务的AscendCL stream流。
  • 返回值：

    返回aclnnStatus状态码，具体参见aclnn返回码。

aclnnLayerNormWithImplModeGetWorkspaceSize

  • 接口定义：

    aclnnStatus aclnnLayerNormWithImplModeGetWorkspaceSize(const aclTensor *input, const aclIntArray *normalizedShape, const aclTensor *weightOptional, const aclTensor *biasOptional, double eps, aclTensor *out, aclTensor *meanOutOptional, aclTensor *rstdOutOptional, int32_t implMode, uint64_t *workspaceSize, aclOpExecutor **executor)

  • 参数说明：
    • input：Device侧的aclTensor，公式中的x，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape长度≥normalizedShape长度，且与normalizedShape右对齐时对应的维度shape相等。支持非连续的Tensor，数据格式支持ND。
    • normalizedShape：Host侧的aclIntArray，表示需要进行norm计算的shape，数据类型支持INT64。其长度小于等于input的shape长度，与input的shape右对齐时的维度shape相等。
    • weightOptional：Device侧的aclTensor，公式中的w，可选参数。weightOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与normalizedShape相等。支持非连续的Tensor，数据格式支持ND。weightOptional为空时，需要构造一个shape与normalizedShape相等、数据全为1的张量。
    • biasOptional：Device侧的aclTensor，公式中的bias，可选参数。biasOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与normalizedShape相等。支持非连续的Tensor，数据格式支持ND。biasOptional为空时，需要构造一个shape与normalizedShape相等、数据全为0的张量。
    • eps：Host侧的浮点数，公式中的eps，用于规避除零计算，数据类型为DOUBLE，需要是可转换成与input相同的数据类型。
    • out：Device侧的aclTensor，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与input的shape相等。支持非连续的Tensor，数据格式支持ND。
    • meanOutOptional：Device侧的aclTensor，可选参数。meanOutOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与rstdOutOptional的shape相等。支持非连续的Tensor，数据格式支持ND。
    • rstdOutOptional：Device侧的aclTensor，可选参数。rstdOutOptional非空时，数据类型支持FLOAT、FLOAT16、BFLOAT16（仅Atlas A2训练系列产品支持）。shape与meanOutOptional的shape相等。支持非连续的Tensor，数据格式支持ND。
    • implMode：Host侧的整型，精度模式，用于指定kernel选择对应的计算模式。数据类型支持INT32。目前支持3种取值，其中0表示高精度、1表示高性能、2表示保持FLOAT16计算模式。
    • workspaceSize：返回用户需要在Device侧申请的workspace大小。
    • executor：返回op执行器，包含了算子计算流程。
  • 返回值：

    返回aclnnStatus状态码，具体参见aclnn返回码。

    第一段接口完成入参校验，出现以下场景时报错：

    • 返回161001（ACLNN_ERR_PARAM_NULLPTR）：传入的input、normalizedShape或out为空指针。
    • 返回161002（ACLNN_ERR_PARAM_INVALID）：
      • input、normalizedShape、weightOptional非空时、biasOptional非空时、out、meanOutOptional非空时或rstdOutOptional非空时的shape超过8维。
      • input、weightOptional非空时、biasOptional非空时、out、meanOutOptional非空时或rstdOutOptional非空时的数据类型不在支持的范围内。
      • normalizedShape维度小于1维。
      • weightOptional非空且shape与normalizedShape不相等。
      • biasOptional非空且shape与normalizedShape不相等。
      • input的维度小于normalizedShape的维度。
      • input的shape与normalizedShape右对齐时对应维度shape不相等。
      • implMode的取值不在0、1、2取值范围内。

aclnnLayerNormWithImplMode

  • 接口定义：

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

  • 参数说明：
    • workspace：在Device侧申请的workspace内存起址。
    • workspaceSize：在Device侧申请的workspace大小，由第一段接口aclnnLayerNormWithImplModeGetWorkspaceSize获取。
    • executor：op执行器，包含了算子计算流程。
    • stream：指定执行任务的AscendCL stream流。
  • 返回值：

    返回aclnnStatus状态码，具体参见aclnn返回码。

调用示例

  • aclnnLayerNorm的调用示例代码如下，仅供参考：
    #include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_layer_norm.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, aclrtContext* context, 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 = aclrtCreateContext(context, deviceId);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateContext failed. ERROR: %d\n", ret); return ret);
  ret = aclrtSetCurrentContext(*context);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetCurrentContext 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 CreateAclTensorMem(const std::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr) {
  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);
  return 0;
}

template <typename T>
void aclCreateTensorP(const std::vector<T>& shape, void** deviceAddr, aclDataType dataType, aclTensor** tensor) {
  // 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);
}

template <typename T>
int CreateAclIntArrayMem(const std::vector<T>& hostData, void** deviceAddr) {
  auto size = GetShapeSize(hostData) * 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);
  return 0;
}

template <typename T>
void aclCreateIntArrayP(const std::vector<T>& hostData, aclIntArray** intArray) {
  // 调用接口创建aclIntArray
  *intArray = aclCreateIntArray(hostData.data(), hostData.size());
}

int main() {
  // 1.（固定写法）device/context/stream初始化，参考AscendCL对外接口列表
  // 根据自己的实际device填写deviceId
  int32_t deviceId = 0;
  aclrtContext context;
  aclrtStream stream;
  auto ret = Init(deviceId, &context, &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> xShape = {1, 800, 5120};
  std::vector<int64_t> normShape = {5120};
  std::vector<int64_t> meanShape = {1, 800, 1};
  void* xDeviceAddr = nullptr;
  void* normShapeAddr = nullptr;
  void* weightDeviceAddr = nullptr;
  void* biasDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  void* meanDeviceAddr = nullptr;
  void* rstdDeviceAddr = nullptr;
  aclTensor* x = nullptr;
  aclIntArray* norm = nullptr;
  aclTensor* weight = nullptr;
  aclTensor* bias = nullptr;
  aclTensor* out = nullptr;
  aclTensor* mean = nullptr;
  aclTensor* rstd = nullptr;
  std::vector<uint16_t> xHostData(4096000, 2.0);
  std::vector<int64_t> normData = {5120};
  std::vector<uint16_t> weightHostData(5120, 1.0);
  std::vector<uint16_t> biasHostData(5120, 0.0);
  std::vector<uint16_t> outHostData(4096000, 0.0);
  std::vector<uint16_t> meanHostData(800, 0.0);
  std::vector<uint16_t> rstdHostData(800, 0.0);
  double eps = 1e-5;

  // 创建x aclTensor
  ret = CreateAclTensorMem(xHostData, xShape, &xDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建normalizedShape aclIntArray
  ret = CreateAclIntArrayMem(normData, &normShapeAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建weight aclTensor
  ret = CreateAclTensorMem(weightHostData, normShape, &weightDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建bias aclTensor
  ret = CreateAclTensorMem(biasHostData, normShape, &biasDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建out aclTensor
  ret = CreateAclTensorMem(outHostData, xShape, &outDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建mean aclTensor
  ret = CreateAclTensorMem(meanHostData, meanShape, &meanDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建rstd aclTensor
  ret = CreateAclTensorMem(rstdHostData, meanShape, &rstdDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  aclCreateTensorP(xShape, &xDeviceAddr, aclDataType::ACL_FLOAT16, &x);
  aclCreateIntArrayP(normData, &norm);
  aclCreateTensorP(normShape, &weightDeviceAddr, aclDataType::ACL_FLOAT16, &weight);
  aclCreateTensorP(normShape, &biasDeviceAddr, aclDataType::ACL_FLOAT16, &bias);
  aclCreateTensorP(xShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
  aclCreateTensorP(meanShape, &meanDeviceAddr, aclDataType::ACL_FLOAT16, &mean);
  aclCreateTensorP(meanShape, &rstdDeviceAddr, aclDataType::ACL_FLOAT16, &rstd);

  // 3. 调用CANN算子库API，需要修改为具体的API名称
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // 调用aclnnLayerNorm第一段接口
  ret = aclnnLayerNormGetWorkspaceSize(x, norm, weight, bias, eps, out, mean, rstd, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnLayerNormGetWorkspaceSize 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aclnnLayerNorm第二段接口
  ret = aclnnLayerNorm(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnLayerNorm 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，需要根据具体API的接口定义修改
  auto size = GetShapeSize(xShape);
  std::vector<float> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
                    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("out result[%ld] is: %f\n", i, resultData[i]);
  }
  auto size1 = GetShapeSize(meanShape);
  std::vector<float> resultData1(size1, 0);
  ret = aclrtMemcpy(resultData1.data(), resultData1.size() * sizeof(resultData1[0]), meanDeviceAddr,
                    size1 * sizeof(resultData1[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 < size1; i++) {
    LOG_PRINT("mean result[%ld] is: %f\n", i, resultData1[i]);
  }
  auto size2 = GetShapeSize(meanShape);
  std::vector<float> resultData2(size2, 0);
  ret = aclrtMemcpy(resultData2.data(), resultData2.size() * sizeof(resultData2[0]), rstdDeviceAddr,
                    size2 * sizeof(resultData2[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 < size2; i++) {
    LOG_PRINT("rstd result[%ld] is: %f\n", i, resultData2[i]);
  }

  // 6. 释放aclTensor和aclIntArray，需要根据具体API的接口定义修改
  aclDestroyTensor(x);
  aclDestroyIntArray(norm);
  aclDestroyTensor(weight);
  aclDestroyTensor(bias);
  aclDestroyTensor(out);
  aclDestroyTensor(mean);
  aclDestroyTensor(rstd);
  return 0;
}
  • aclnnLayerNormWithImplMode调用示例代码如下，仅供参考：
    #include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_layer_norm.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, aclrtContext* context, 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 = aclrtCreateContext(context, deviceId);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateContext failed. ERROR: %d\n", ret); return ret);
  ret = aclrtSetCurrentContext(*context);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetCurrentContext 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 CreateAclTensorMem(const std::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr) {
  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);
  return 0;
}

template <typename T>
void aclCreateTensorP(const std::vector<T>& shape, void** deviceAddr, aclDataType dataType, aclTensor** tensor) {
  // 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);
}

template <typename T>
int CreateAclIntArrayMem(const std::vector<T>& hostData, void** deviceAddr) {
  auto size = GetShapeSize(hostData) * 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);
  return 0;
}

template <typename T>
void aclCreateIntArrayP(const std::vector<T>& hostData, aclIntArray** intArray) {
  // 调用接口创建aclIntArray
  *intArray = aclCreateIntArray(hostData.data(), hostData.size());
}

int main() {
  // 1.（固定写法）device/context/stream初始化，参考AscendCL对外接口列表
  // 根据自己的实际device填写deviceId
  int32_t deviceId = 0;
  aclrtContext context;
  aclrtStream stream;
  auto ret = Init(deviceId, &context, &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> xShape = {1, 800, 5120};
  std::vector<int64_t> normShape = {5120};
  std::vector<int64_t> meanShape = {1, 800, 1};
  void* xDeviceAddr = nullptr;
  void* normShapeAddr = nullptr;
  void* weightDeviceAddr = nullptr;
  void* biasDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  void* meanDeviceAddr = nullptr;
  void* rstdDeviceAddr = nullptr;
  aclTensor* x = nullptr;
  aclIntArray* norm = nullptr;
  aclTensor* weight = nullptr;
  aclTensor* bias = nullptr;
  aclTensor* out = nullptr;
  aclTensor* mean = nullptr;
  aclTensor* rstd = nullptr;
  std::vector<uint16_t> xHostData(4096000, 2.0);
  std::vector<int64_t> normData = {5120};
  std::vector<uint16_t> weightHostData(5120, 1.0);
  std::vector<uint16_t> biasHostData(5120, 0.0);
  std::vector<uint16_t> outHostData(4096000, 0.0);
  std::vector<uint16_t> meanHostData(800, 0.0);
  std::vector<uint16_t> rstdHostData(800, 0.0);
  double eps = 1e-5;
  int32_t implMode = 2;

  // 创建x aclTensor
  ret = CreateAclTensorMem(xHostData, xShape, &xDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建normalizedShape aclIntArray
  ret = CreateAclIntArrayMem(normData, &normShapeAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建weight aclTensor
  ret = CreateAclTensorMem(weightHostData, normShape, &weightDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建bias aclTensor
  ret = CreateAclTensorMem(biasHostData, normShape, &biasDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建out aclTensor
  ret = CreateAclTensorMem(outHostData, xShape, &outDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建mean aclTensor
  ret = CreateAclTensorMem(meanHostData, meanShape, &meanDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建rstd aclTensor
  ret = CreateAclTensorMem(rstdHostData, meanShape, &rstdDeviceAddr);
  CHECK_RET(ret == ACL_SUCCESS, return ret);

  aclCreateTensorP(xShape, &xDeviceAddr, aclDataType::ACL_FLOAT16, &x);
  aclCreateIntArrayP(normData, &norm);
  aclCreateTensorP(normShape, &weightDeviceAddr, aclDataType::ACL_FLOAT16, &weight);
  aclCreateTensorP(normShape, &biasDeviceAddr, aclDataType::ACL_FLOAT16, &bias);
  aclCreateTensorP(xShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
  aclCreateTensorP(meanShape, &meanDeviceAddr, aclDataType::ACL_FLOAT16, &mean);
  aclCreateTensorP(meanShape, &rstdDeviceAddr, aclDataType::ACL_FLOAT16, &rstd);

  // 3. 调用CANN算子库API，需要修改为具体的API名称
  uint64_t workspaceSize = 0;
  aclOpExecutor* executor;
  // 调用aclnnLayerNormWithImplMode第一段接口
  ret = aclnnLayerNormWithImplModeGetWorkspaceSize(x, norm, weight, bias, eps, out, mean, rstd, implMode,
                                                   &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS,
            LOG_PRINT("aclnnLayerNormWithImplModeGetWorkspaceSize 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aclnnLayerNormWithImplMode第二段接口
  ret = aclnnLayerNormWithImplMode(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnLayerNormWithImplMode 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，需要根据具体API的接口定义修改
  auto size = GetShapeSize(xShape);
  std::vector<float> resultData(size, 0);
  ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), outDeviceAddr,
                    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("out result[%ld] is: %f\n", i, resultData[i]);
  }
  auto size1 = GetShapeSize(meanShape);
  std::vector<float> resultData1(size1, 0);
  ret = aclrtMemcpy(resultData1.data(), resultData1.size() * sizeof(resultData1[0]), meanDeviceAddr,
                    size1 * sizeof(resultData1[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 < size1; i++) {
    LOG_PRINT("mean result[%ld] is: %f\n", i, resultData1[i]);
  }
  auto size2 = GetShapeSize(meanShape);
  std::vector<float> resultData2(size2, 0);
  ret = aclrtMemcpy(resultData2.data(), resultData2.size() * sizeof(resultData2[0]), rstdDeviceAddr,
                    size2 * sizeof(resultData2[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 < size2; i++) {
    LOG_PRINT("rstd result[%ld] is: %f\n", i, resultData2[i]);
  }

  // 6. 释放aclTensor和aclIntArray，需要根据具体API的接口定义修改
  aclDestroyTensor(x);
  aclDestroyIntArray(norm);
  aclDestroyTensor(weight);
  aclDestroyTensor(bias);
  aclDestroyTensor(out);
  aclDestroyTensor(mean);
  aclDestroyTensor(rstd);
  return 0;
}
父主题： NN类算子接口