aclnnQuantMatmulV3

支持的产品型号

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

接口原型

每个算子分为两段式接口，必须先调用“aclnnQuantMatmulV3GetWorkspaceSize”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“aclnnQuantMatmulV3”接口执行计算。

• aclnnStatus aclnnQuantMatmulV3GetWorkspaceSize(const aclTensor* x1, const aclTensor* x2, const aclTensor* scale, const aclTensor* offset, const aclTensor* bias, bool transposeX1, bool transposeX2, const aclTensor* out, uint64_t* workspaceSize, aclOpExecutor** executor)

• aclnnStatus aclnnQuantMatmulV3(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream)

功能描述

• 算子功能：完成量化的矩阵乘计算，最小支持输入维度为2维，最大支持输入维度为6维。相似接口有aclnnMm（仅支持2维Tensor作为输入的矩阵乘）和aclnnBatchMatMul（仅支持三维的矩阵乘，其中第一维是Batch维度）。

• 计算公式：

  • 无bias：

    $$out = x1@x2 * scale + offset$$

  • bias int32：

    $$out = (x1@x2 + bias) * scale + offset$$

  • bias bfloat16/float32（此场景无offset）:

    $$out = x1@x2 * scale + bias$$

aclnnQuantMatmulV3GetWorkspaceSize

• 参数说明：

  • x1（aclTensor*，计算输入）：公式中的输入x1，数据类型支持INT8、INT32、INT4，支持最后两根轴转置情况下的非连续tensor，其他场景的非连续的Tensor不支持，数据格式支持ND，shape最少是2维，最多是6维，在transposeX1为false情况下各个维度表示：（batch，m，k），batch可不存在。当数据类型为INT32、INT4时，为INT4量化场景，当前仅支持Atlas A2训练系列产品/Atlas 800I A2推理产品，2-6维ND格式，transposeX1为false情况。其中当x1数据类型为INT4时，维度表示：(batch, m, k)，要求k为偶数，当x1数据类型为INT32时，每个INT32数据存放8个INT4数据，对应维度表示：(batch, m, k // 8)，要求k为8的倍数。

  • x2（aclTensor*，计算输入）：公式中的输入x2，数据类型支持INT8、INT32、INT4，支持最后两根轴转置情况下的非连续tensor，其他场景的非连续的Tensor不支持，数据格式支持ND格式和昇腾亲和数据排布格式。当数据类型为INT32、INT4时，为INT4量化场景，当前仅支持Atlas A2训练系列产品/Atlas 800I A2推理产品，2维ND格式。

    • ND格式下，shape最少是2维，最多是6维，在transposeX2为false情况下各个维度表示：（batch，k，n），batch可不存在，其中k与x1的shape中的k一致。
    • 昇腾亲和数据排布格式下，shape最少是4维，最多是8维。在transposeX2为true情况下各个维度表示：（batch，k1，n1, n0, k0），batch可不存在，其中k0 = 32, n0 = 16, x1 shape中的k和x2 shape中的k1需要满足以下关系：ceilDiv(k,32) = k1。在transposeX2为false情况下各个维度表示：（batch，n1，k1, k0, n0）,batch可不存在，其中k0 = 16, n0 = 32, x1 shape中的k和x2 shape中的k1需要满足以下关系：ceilDiv(k,16) = k1。 可使用aclnnCalculateMatmulWeightSizeV2接口以及aclnnTransMatmulWeight接口完成输入Format从ND到昇腾亲和数据排布格式的转换。（当输入x2为昇腾亲和数据排布格式时，当前Atlas 推理系列产品上QuantBatchMatmulV3算子不支持transposeX2为false的场景）
    • 数据类型为INT4时，在transposeX2为true情况下各个维度表示：（n, k），要求k为偶数；在transposeX2为false情况下各个维度表示：（k, n），要求n为偶数。数据类型为INT32时，每个INT32数据存放8个INT4数据，在transposeX2为true情况下各个维度表示：(n, k // 8)，要求k为8的倍数；在transposeX2为false情况下各个维度表示：(k, n // 8)，要求n为8的倍数。 可使用aclnnConvertWeightToINT4Pack接口完成x2从INT32（1个int32在0~3bit位存储1个int4）到INT32（1个int32存储8个int4）或INT4（1个int4表示1个int4）的数据格式转换，具体参见aclnnConvertWeightToINT4Pack接口。

  • scale（aclTensor*，计算输入）：公式中的输入scale，量化参数，数据类型支持UINT64，INT64，FLOAT32，BFLOAT16，数据格式支持ND，shape是1维（t，），t = 1或n，其中n与x2的n一致。

    • 输出为INT8、FLOAT16时，需要提前调用TransQuantParamV2算子的aclnn接口来将scale转成INT64、UINT64数据类型。
    • 输出为BFLOAT16时，直接将BFLOAT16或FLOAT32类型的scale传入本接口，其中BFLOAT16仅Atlas A2训练系列产品/Atlas 800I A2推理产品 支持。

  • offset（aclTensor*，计算输入）：公式中的输入offset，可选量化参数，数据类型支持FLOAT32，数据格式支持ND，shape是1维（t，），t = 1或n，其中n与x2的n一致。

  • bias（aclTensor*，计算输入）：公式中的输入bias，可选参数。数据类型支持INT32，BFLOAT16，FLOAT32，数据格式支持ND，shape支持1维（n，）或3维（batch，1，n），n与x2的n一致。当out的shape为2、4、5、6维或x1和x2为int32/int4时，bias的shape只支持1维(n,)。

  • transposeX1（bool，计算输入）：表示x1的输入shape是否包含transpose，默认是false，若为true，x1的shape表示为（batch，k，m），batch可不存在。x1和x2为int32/int4时，transposeX1仅支持false。

  • transposeX2（bool，计算输入）：表示x2的输入shape是否包含transpose，默认是false，若为true，x2的shape表示为（batch，n，k），batch可不存在。

  • out（aclTensor*，计算输出）：公式中的输出out，数据类型支持FLOAT16，INT8，BFLOAT16，支持非连续的Tensor，数据格式支持ND，shape最少是2维，最多是6维，（batch，m，n），batch可不存在，支持x1与x2的batch维度broadcast，输出batch与broadcast之后的batch一致，m与x1的m一致，n与x2的n一致。其中BFLOAT16仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持。

  • 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。

  第一段接口完成入参校验，出现以下场景时报错：
  - 161001(ACLNN_ERR_PARAM_NULLPTR)：
  1. 传入的x1、x2、scale或out是空指针。
  - 161002(ACLNN_ERR_PARAM_INVALID)：
  1. x1、x2、bias、scale、offset或out的数据类型和数据格式不在支持的范围之内。
  2. x1、x2、bias、scale、offset或out的shape不满足校验条件。
  3. x1、x2、bias、scale、offset或out是空tensor。
  4. x1与x2的最后一维大小超过65535，x1的最后一维指transposeX1为true时的m或transposeX1为false时的k，x2的最后一维指transposeX2为true时的k或transposeX2为false时的n。

aclnnQuantMatmulV3

• 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aclnnQuantMatmulV3GetWorkspaceSize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

输入和输出支持以下数据类型组合：

x1	x2	scale	offset	bias	out	产品类型
int8	int8	uint64/int64	null	int32	float16	Atlas 推理系列产品/Atlas A2训练系列产品/Atlas 800I A2推理产品
int8	int8	uint64/int64	float32	int32	int8	Atlas 推理系列产品/Atlas A2训练系列产品/Atlas 800I A2推理产品
int8	int8	float32/bfloat16	null	int32/bfloat16/float32	bfloat16	Atlas A2训练系列产品/Atlas 800I A2推理产品
int4/int32	int4/int32	uint64	null	int32	float16	Atlas A2训练系列产品/Atlas 800I A2推理产品

调用示例

示例代码如下（以Atlas A2训练系列产品/Atlas 800I A2推理产品为例），仅供参考，具体编译和执行过程请参考编译与运行样例。

#include <memory>
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_quant_matmul_v3.h"
#include "aclnnop/aclnn_trans_quant_param_v2.h"

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

#define CHECK_FREE_RET(cond, return_expr) \
  do {                               \
    if (!(cond)) {                   \
      Finalize(deviceId, stream);\
      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;
}

void Finalize(int32_t deviceId, aclrtStream stream) {
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
}

int aclnnQuantMatmulV3Test(int32_t deviceId, 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> x1Shape = {5, 2};
  std::vector<int64_t> x2Shape = {2, 3};
  std::vector<int64_t> biasShape = {3};
  std::vector<int64_t> offsetShape = {3};
  std::vector<int64_t> scaleShape = {3};
  std::vector<int64_t> outShape = {5, 3};
  void* x1DeviceAddr = nullptr;
  void* x2DeviceAddr = nullptr;
  void* scaleDeviceAddr = nullptr;
  void* quantParamDeviceAddr = nullptr;
  void* offsetDeviceAddr = nullptr;
  void* biasDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* x1 = nullptr;
  aclTensor* x2 = nullptr;
  aclTensor* bias = nullptr;
  aclTensor* scale = nullptr;
  aclTensor* quantParam = nullptr;
  aclTensor* offset = nullptr;
  aclTensor* out = nullptr;
  std::vector<int8_t> x1HostData = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
  std::vector<int8_t> x2HostData = {1, 1, 1, 1, 1, 1};
  std::vector<int32_t> biasHostData = {1, 1, 1};
  std::vector<float> scaleHostData = {1, 1, 1};
  std::vector<float> offsetHostData = {1, 1, 1};
  std::vector<uint16_t> outHostData = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; // 实际上是float16半精度方式
  // 创建x1 aclTensor
  ret = CreateAclTensor(x1HostData, x1Shape, &x1DeviceAddr, aclDataType::ACL_INT8, &x1);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x1TensorPtr(x1, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x1DeviceAddrPtr(x1DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建x2 aclTensor
  ret = CreateAclTensor(x2HostData, x2Shape, &x2DeviceAddr, aclDataType::ACL_INT8, &x2);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x2TensorPtr(x2, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x2DeviceAddrPtr(x2DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建scale aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &scaleDeviceAddr, aclDataType::ACL_FLOAT, &scale);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> scaleTensorPtr(scale, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> scaleDeviceAddrPtr(scaleDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建quantParam aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &quantParamDeviceAddr, aclDataType::ACL_UINT64, &quantParam);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> quantParamTensorPtr(quantParam, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> quantParamDeviceAddrPtr(quantParamDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建offset aclTensor
  ret = CreateAclTensor(offsetHostData, offsetShape, &offsetDeviceAddr, aclDataType::ACL_FLOAT, &offset);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> offsetTensorPtr(offset, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> offsetDeviceAddrPtr(offsetDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建bias aclTensor
  ret = CreateAclTensor(biasHostData, biasShape, &biasDeviceAddr, aclDataType::ACL_INT32, &bias);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> biasTensorPtr(bias, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> biasDeviceAddrPtr(biasDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建out aclTensor
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> outTensorPtr(out, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> outDeviceAddrPtr(outDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  bool transposeX1 = false;
  bool transposeX2 = false;

  // 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;

  // FLOAT数据类型的scale需要提前调用TransQuantParamV2算子的aclnn接口
  // 调用aclnnaclnnTransQuantParamV2第一段接口
  ret = aclnnTransQuantParamV2GetWorkspaceSize(scale, offset, quantParam, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2GetWorkspaceSize 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;
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV2(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV2.reset(workspaceAddr);
  }
  // 调用aclnnTransQuantParamV2第二段接口
  ret = aclnnTransQuantParamV2(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2 failed. ERROR: %d\n", ret); return ret);

  // 调用aclnnQuantMatmulV3第一段接口
  ret = aclnnQuantMatmulV3GetWorkspaceSize(x1, x2, quantParam, nullptr, bias, transposeX1, transposeX2, out, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  workspaceAddr = nullptr;
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV3(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV3.reset(workspaceAddr);
  }

  // 调用aclnnQuantMatmulV3第二段接口
  ret = aclnnQuantMatmulV3(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3 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(outShape);
  std::vector<uint16_t> resultData(size, 0); // C语言中无法直接打印fp16的数据，需要用uint16读出来，自行通过二进制转成fp16
  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("result[%ld] is: %u\n", i, resultData[i]);
  }
  return ACL_SUCCESS;
}

int main() {
  // 1. （固定写法）device/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;
  aclrtStream stream;
  auto ret = aclnnQuantMatmulV3Test(deviceId, stream);
  CHECK_FREE_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3Test failed. ERROR: %d\n", ret); return ret);

  Finalize(deviceId, stream);
  return 0;
}

Atlas A2训练系列产品/Atlas 800I A2推理产品x2为昇腾亲和数据排布格式场景的示例代码如下(transposeX2=false)，仅供参考，具体编译和执行过程请参考编译与运行样例。

#include <memory>
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_quant_matmul_v3.h"
#include "aclnnop/aclnn_trans_matmul_weight.h"
#include "aclnnop/aclnn_permute.h"
#include "aclnnop/aclnn_trans_quant_param_v2.h"

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

#define CHECK_FREE_RET(cond, return_expr) \
  do {                               \
    if (!(cond)) {                   \
      Finalize(deviceId, stream);\
      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;
}

void Finalize(int32_t deviceId, aclrtStream stream) {
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
}

template <typename T>
int CreateAclTensorX2(const std::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr,
                          aclDataType dataType, aclTensor** tensor) {
  auto size = static_cast<uint64_t>(GetShapeSize(shape));

  const aclIntArray* mat2Size = aclCreateIntArray(shape.data(), shape.size());
  auto ret = aclnnCalculateMatmulWeightSizeV2(mat2Size, dataType, &size);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnCalculateMatmulWeightSizeV2 failed. ERROR: %d\n", ret); return ret);
  size *= sizeof(T);

  // 调用aclrtMalloc申请device侧内存
  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];
  }

  std::vector<int64_t> storageShape;
  storageShape.push_back(GetShapeSize(shape));

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

int aclnnQuantMatmulV3Test(int32_t deviceId, 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> x1Shape = {5, 2};
  std::vector<int64_t> x2Shape = {2, 3};
  std::vector<int64_t> biasShape = {3};
  std::vector<int64_t> offsetShape = {3};
  std::vector<int64_t> scaleShape = {3};
  std::vector<int64_t> outShape = {5, 3};
  void* x1DeviceAddr = nullptr;
  void* x2DeviceAddr = nullptr;
  void* scaleDeviceAddr = nullptr;
  void* quantParamDeviceAddr = nullptr;
  void* offsetDeviceAddr = nullptr;
  void* biasDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* x1 = nullptr;
  aclTensor* x2 = nullptr;
  aclTensor* bias = nullptr;
  aclTensor* scale = nullptr;
  aclTensor* quantParam = nullptr;
  aclTensor* offset = nullptr;
  aclTensor* out = nullptr;
  std::vector<int8_t> x1HostData = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
  std::vector<int8_t> x2HostData = {1, 1, 1, 1, 1, 1};
  std::vector<int32_t> biasHostData = {1, 1, 1};
  std::vector<float> scaleHostData = {1, 1, 1};
  std::vector<float> offsetHostData = {1, 1, 1};
  std::vector<uint16_t> outHostData = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; // 实际上是float16半精度方式
  // 创建x1 aclTensor
  ret = CreateAclTensor(x1HostData, x1Shape, &x1DeviceAddr, aclDataType::ACL_INT8, &x1);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x1TensorPtr(x1, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x1DeviceAddrPtr(x1DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建昇腾亲和数据排布格式的x2 aclTensor
  ret = CreateAclTensorX2(x2HostData, x2Shape, &x2DeviceAddr, aclDataType::ACL_INT8, &x2);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x2HPTensorPtr(x2, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x2HPDeviceAddrPtr(x2DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建scale aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &scaleDeviceAddr, aclDataType::ACL_FLOAT, &scale);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> scaleTensorPtr(scale, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> scaleDeviceAddrPtr(scaleDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建quantParam aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &quantParamDeviceAddr, aclDataType::ACL_UINT64, &quantParam);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> quantParamTensorPtr(quantParam, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> quantParamDeviceAddrPtr(quantParamDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建offset aclTensor
  ret = CreateAclTensor(offsetHostData, offsetShape, &offsetDeviceAddr, aclDataType::ACL_FLOAT, &offset);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> offsetTensorPtr(offset, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> offsetDeviceAddrPtr(offsetDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建bias aclTensor
  ret = CreateAclTensor(biasHostData, biasShape, &biasDeviceAddr, aclDataType::ACL_INT32, &bias);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> biasTensorPtr(bias, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> biasDeviceAddrPtr(biasDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建out aclTensor
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> outTensorPtr(out, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> outDeviceAddrPtr(outDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  bool transposeX1 = false;
  bool transposeX2 = false;

  // 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;
  void* workspaceAddr = nullptr;
  // 调用aclnnTransMatmulWeight第一段接口
  ret = aclnnTransMatmulWeightGetWorkspaceSize(x2, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransMatmulWeightGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);

  // 根据第一段接口计算出的workspaceSize申请device内存
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrTrans(nullptr, aclrtFree);
  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);
    workspaceAddrPtrTrans.reset(workspaceAddr);
  }
  // 调用aclnnTransMatmulWeight第二段接口
  ret = aclnnTransMatmulWeight(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransMatmulWeight failed. ERROR: %d\n", ret); return ret);

  // FLOAT数据类型的scale需要提前调用TransQuantParamV2算子的aclnn接口
  // 调用aclnnaclnnTransQuantParamV2第一段接口
  ret = aclnnTransQuantParamV2GetWorkspaceSize(scale, offset, quantParam, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  workspaceAddr = nullptr;
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV2(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV2.reset(workspaceAddr);
  }
  // 调用aclnnTransQuantParamV2第二段接口
  ret = aclnnTransQuantParamV2(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2 failed. ERROR: %d\n", ret); return ret);

  // 调用aclnnQuantMatmulV3第一段接口
  workspaceSize = 0;
  ret = aclnnQuantMatmulV3GetWorkspaceSize(x1, x2, quantParam, nullptr, bias, transposeX1, transposeX2, out, &workspaceSize, &executor);

  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存

  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV3(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV3.reset(workspaceAddr);
  }
  // 调用aclnnQuantMatmulV3第二段接口
  ret = aclnnQuantMatmulV3(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3 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(outShape);
  std::vector<uint16_t> resultData(size, 0); // C语言中无法直接打印fp16的数据，需要用uint16读出来，自行通过二进制转成fp16
  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("result[%ld] is: %u\n", i, resultData[i]);
  }
  return ACL_SUCCESS;
}

int main() {
  // 1. （固定写法）device/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;
  aclrtStream stream;
  auto ret = aclnnQuantMatmulV3Test(deviceId, stream);
  CHECK_FREE_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3Test failed. ERROR: %d\n", ret); return ret);

  Finalize(deviceId, stream);
  return 0;
}

Atlas 推理系列产品x2为昇腾亲和数据排布格式场景的示例代码如下(transposeX2=true)，仅供参考，具体编译和执行过程请参考编译与运行样例。

#include <memory>
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_quant_matmul_v3.h"
#include "aclnnop/aclnn_trans_matmul_weight.h"
#include "aclnnop/aclnn_permute.h"
#include "aclnnop/aclnn_trans_quant_param_v2.h"

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

#define CHECK_FREE_RET(cond, return_expr) \
  do {                               \
    if (!(cond)) {                   \
      Finalize(deviceId, stream);\
      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;
}

void Finalize(int32_t deviceId, aclrtStream stream) {
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
}

template <typename T>
int CreateAclTensorX2(const std::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr,
                          aclDataType dataType, aclTensor** tensor) {
  auto size = static_cast<uint64_t>(GetShapeSize(shape));

  const aclIntArray* mat2Size = aclCreateIntArray(shape.data(), shape.size());
  auto ret = aclnnCalculateMatmulWeightSizeV2(mat2Size, dataType, &size);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnCalculateMatmulWeightSizeV2 failed. ERROR: %d\n", ret); return ret);
  size *= sizeof(T);

  // 调用aclrtMalloc申请device侧内存
  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];
  }

  std::vector<int64_t> storageShape;
  storageShape.push_back(GetShapeSize(shape));

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

int aclnnQuantMatmulV3Test(int32_t deviceId, 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> x1Shape = {5, 2};
  std::vector<int64_t> x2Shape = {2, 3};
  std::vector<int64_t> x2TransposedShape = {3, 2};
  std::vector<int64_t> biasShape = {3};
  std::vector<int64_t> offsetShape = {3};
  std::vector<int64_t> scaleShape = {3};
  std::vector<int64_t> outShape = {5, 3};
  void* x1DeviceAddr = nullptr;
  void* x2DeviceAddr = nullptr;
  void* x2TransposedDeviceAddr = nullptr;
  void* scaleDeviceAddr = nullptr;
  void* quantParamDeviceAddr = nullptr;
  void* offsetDeviceAddr = nullptr;
  void* biasDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* x1 = nullptr;
  aclTensor* x2 = nullptr;
  aclTensor* x2Transposed = nullptr;
  aclTensor* bias = nullptr;
  aclTensor* scale = nullptr;
  aclTensor* quantParam = nullptr;
  aclTensor* offset = nullptr;
  aclTensor* out = nullptr;
  std::vector<int8_t> x1HostData = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1};
  std::vector<int8_t> x2HostData = {1, 1, 1, 1, 1, 1};
  std::vector<int8_t> x2TransposedHostData = {1, 1, 1, 1, 1, 1};
  std::vector<int32_t> biasHostData = {1, 1, 1};
  std::vector<float> scaleHostData = {1, 1, 1};
  std::vector<float> offsetHostData = {1, 1, 1};
  std::vector<uint16_t> outHostData = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; // 实际上是float16半精度方式
  // 创建x1 aclTensor
  ret = CreateAclTensor(x1HostData, x1Shape, &x1DeviceAddr, aclDataType::ACL_INT8, &x1);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x1TensorPtr(x1, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x1DeviceAddrPtr(x1DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建昇腾亲和数据排布格式的x2 aclTensor
  ret = CreateAclTensorX2(x2HostData, x2Shape, &x2DeviceAddr, aclDataType::ACL_INT8, &x2);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x2HPTensorPtr(x2, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x2HPDeviceAddrPtr(x2DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  //创建昇腾亲和数据排布格式的x2Transposed aclTensor
  ret = CreateAclTensorX2(x2TransposedHostData, x2TransposedShape, &x2TransposedDeviceAddr, aclDataType::ACL_INT8, &x2Transposed);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x2TransposedHPTensorPtr(x2Transposed, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x2TransposedHPDeviceAddrPtr(x2TransposedDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建scale aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &scaleDeviceAddr, aclDataType::ACL_FLOAT, &scale);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> scaleTensorPtr(scale, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> scaleDeviceAddrPtr(scaleDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建quantParam aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &quantParamDeviceAddr, aclDataType::ACL_UINT64, &quantParam);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> quantParamTensorPtr(quantParam, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> quantParamDeviceAddrPtr(quantParamDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建offset aclTensor
  ret = CreateAclTensor(offsetHostData, offsetShape, &offsetDeviceAddr, aclDataType::ACL_FLOAT, &offset);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> offsetTensorPtr(offset, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> offsetDeviceAddrPtr(offsetDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建bias aclTensor
  ret = CreateAclTensor(biasHostData, biasShape, &biasDeviceAddr, aclDataType::ACL_INT32, &bias);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> biasTensorPtr(bias, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> biasDeviceAddrPtr(biasDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建out aclTensor
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> outTensorPtr(out, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> outDeviceAddrPtr(outDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  bool transposeX1 = false;
  bool transposeX2 = true;

  // 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;
  void* workspaceAddr = nullptr;

  // x2的shpe需要transpose成nk格式，再进行transdata
  std::vector<int64_t> dimsData = {1, 0};
  // 创建dims aclIntArray
  aclIntArray *dims = aclCreateIntArray(dimsData.data(), dimsData.size());
  // 调用aclnnPermute第一段接口
  ret = aclnnPermuteGetWorkspaceSize(x2, dims, x2Transposed, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnPermuteGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrPermute(nullptr, aclrtFree);
  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);
    workspaceAddrPtrPermute.reset(workspaceAddr);
  }
  // 调用aclnnPermute第二段接口
  ret = aclnnPermute(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnPermuteGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);

  workspaceSize = 0;
  // 调用aclnnTransMatmulWeight第一段接口
  ret = aclnnTransMatmulWeightGetWorkspaceSize(x2Transposed, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransMatmulWeightGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);

  // 根据第一段接口计算出的workspaceSize申请device内存
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrTrans(nullptr, aclrtFree);
  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);
    workspaceAddrPtrTrans.reset(workspaceAddr);
  }
  // 调用aclnnTransMatmulWeight第二段接口
  ret = aclnnTransMatmulWeight(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransMatmulWeight failed. ERROR: %d\n", ret); return ret);

  // FLOAT数据类型的scale需要提前调用TransQuantParamV2算子的aclnn接口
  // 调用aclnnaclnnTransQuantParamV2第一段接口
  ret = aclnnTransQuantParamV2GetWorkspaceSize(scale, offset, quantParam, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  workspaceAddr = nullptr;
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV2(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV2.reset(workspaceAddr);
  }
  // 调用aclnnTransQuantParamV2第二段接口
  ret = aclnnTransQuantParamV2(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2 failed. ERROR: %d\n", ret); return ret);

  // 调用aclnnQuantMatmulV3第一段接口
  workspaceSize = 0;
  ret = aclnnQuantMatmulV3GetWorkspaceSize(x1, x2Transposed, quantParam, nullptr, bias, transposeX1, transposeX2, out, &workspaceSize, &executor);

  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存

  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV3(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV3.reset(workspaceAddr);
  }
  // 调用aclnnQuantMatmulV3第二段接口
  ret = aclnnQuantMatmulV3(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3 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(outShape);
  std::vector<uint16_t> resultData(size, 0); // C语言中无法直接打印fp16的数据，需要用uint16读出来，自行通过二进制转成fp16
  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("result[%ld] is: %u\n", i, resultData[i]);
  }
  return ACL_SUCCESS;
}

int main() {
  // 1. （固定写法）device/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;
  aclrtStream stream;
  auto ret = aclnnQuantMatmulV3Test(deviceId, stream);
  CHECK_FREE_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3Test failed. ERROR: %d\n", ret); return ret);

  Finalize(deviceId, stream);
  return 0;
}

Atlas A2训练系列产品/Atlas 800I A2推理产品 INT4量化场景示例代码如下(x1和x2数据类型为int4，transposeX2=false)，仅供参考，具体编译和执行过程请参考编译与运行样例。

#include <memory>
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_quant_matmul_v3.h"
#include "aclnnop/aclnn_trans_quant_param_v2.h"
#include "aclnnop/aclnn_convert_weight_to_int4_pack.h"

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

#define CHECK_FREE_RET(cond, return_expr) \
  do {                               \
    if (!(cond)) {                   \
      Finalize(deviceId, stream);\
      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) {
  // 通过hostData获取申请和拷贝的内存byte数
  auto size = hostData.size() * 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;
}

void Finalize(int32_t deviceId, aclrtStream stream) {
  aclrtDestroyStream(stream);
  aclrtResetDevice(deviceId);
  aclFinalize();
}

int aclnnQuantMatmulV3Test(int32_t deviceId, 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
  int64_t m = 16;
  int64_t k = 8;
  int64_t n = 32;
  aclDataType x1Dtype = aclDataType::ACL_INT4;
  aclDataType x2Int4PackDtype = aclDataType::ACL_INT4;
  std::vector<int64_t> x1Shape = {m, k};
  std::vector<int64_t> x2Shape = {k, n};
  std::vector<int64_t> x2Int4PackShape = {k, n};
  std::vector<int64_t> biasShape = {n};
  std::vector<int64_t> scaleShape = {n};
  std::vector<int64_t> outShape = {m, n};
  void* x1DeviceAddr = nullptr;
  void* x2DeviceAddr = nullptr;
  void* x2Int4PackDeviceAddr = nullptr;
  void* scaleDeviceAddr = nullptr;
  void* quantParamDeviceAddr = nullptr;
  void* offsetDeviceAddr = nullptr;
  void* biasDeviceAddr = nullptr;
  void* outDeviceAddr = nullptr;
  aclTensor* x1 = nullptr;
  aclTensor* x2 = nullptr;
  aclTensor* x2Int4Pack = nullptr;
  aclTensor* bias = nullptr;
  aclTensor* scale = nullptr;
  aclTensor* quantParam = nullptr;
  aclTensor* offset = nullptr;
  aclTensor* out = nullptr;
  std::vector<int8_t> x1HostData(m * k / 2, 17);  // int8: 0001 0001
  std::vector<int8_t> x2HostData(k * n, 1);
  std::vector<int8_t> x2Int4PackHostData(n * k / 2, 1);
  std::vector<int32_t> biasHostData(n, 1);
  std::vector<float> scaleHostData(n, 1);
  std::vector<uint16_t> outHostData(m * n, 1);

  // 创建x1 aclTensor
  ret = CreateAclTensor(x1HostData, x1Shape, &x1DeviceAddr, x1Dtype, &x1);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x1TensorPtr(x1, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x1DeviceAddrPtr(x1DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建x2 aclTensor
  ret = CreateAclTensor(x2HostData, x2Shape, &x2DeviceAddr, aclDataType::ACL_INT32, &x2);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x2TensorPtr(x2, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x2DeviceAddrPtr(x2DeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建x2Int4Pack aclTensor
  ret = CreateAclTensor(x2Int4PackHostData, x2Int4PackShape, &x2Int4PackDeviceAddr, x2Int4PackDtype, &x2Int4Pack);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> x2Int4PackTensorPtr(x2Int4Pack, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> x2Int4PackDeviceAddrPtr(x2Int4PackDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建scale aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &scaleDeviceAddr, aclDataType::ACL_FLOAT, &scale);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> scaleTensorPtr(scale, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> scaleDeviceAddrPtr(scaleDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建quantParam aclTensor
  ret = CreateAclTensor(scaleHostData, scaleShape, &quantParamDeviceAddr, aclDataType::ACL_UINT64, &quantParam);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> quantParamTensorPtr(quantParam, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> quantParamDeviceAddrPtr(quantParamDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建bias aclTensor
  ret = CreateAclTensor(biasHostData, biasShape, &biasDeviceAddr, aclDataType::ACL_INT32, &bias);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> biasTensorPtr(bias, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> biasDeviceAddrPtr(biasDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  // 创建out aclTensor
  ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT16, &out);
  std::unique_ptr<aclTensor, aclnnStatus (*)(const aclTensor *)> outTensorPtr(out, aclDestroyTensor);
  std::unique_ptr<void, aclError (*)(void *)> outDeviceAddrPtr(outDeviceAddr, aclrtFree);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  bool transposeX1 = false;
  bool transposeX2 = false;

  // 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;

  // 可以先调用aclnnConvertWeightToINT4Pack接口来构建x2输入数据
  // 调用aclnnConvertWeightToINT4Pack第一段接口
  ret = aclnnConvertWeightToINT4PackGetWorkspaceSize(x2, x2Int4Pack, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnConvertWeightToINT4PackGetWorkspaceSize 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;
  std::unique_ptr<void, aclError (*)(void *)> workspaceINT4PackAddrPtr(nullptr, aclrtFree);
  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);
    workspaceINT4PackAddrPtr.reset(workspaceAddr);
  }
  // 调用aclnnConvertWeightToINT4Pack第二段接口
  ret = aclnnConvertWeightToINT4Pack(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnConvertWeightToINT4Pack failed. ERROR: %d\n", ret); return ret);

  // FLOAT数据类型的scale需要提前调用TransQuantParamV2算子的aclnn接口
  // 调用aclnnaclnnTransQuantParamV2第一段接口
  ret = aclnnTransQuantParamV2GetWorkspaceSize(scale, offset, quantParam, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  workspaceAddr = nullptr;
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV2(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV2.reset(workspaceAddr);
  }
  // 调用aclnnTransQuantParamV2第二段接口
  ret = aclnnTransQuantParamV2(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTransQuantParamV2 failed. ERROR: %d\n", ret); return ret);

  // 调用aclnnQuantMatmulV3第一段接口
  ret = aclnnQuantMatmulV3GetWorkspaceSize(x1, x2Int4Pack, quantParam, nullptr, bias, transposeX1, transposeX2, out,
                                           &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
  // 根据第一段接口计算出的workspaceSize申请device内存
  workspaceAddr = nullptr;
  std::unique_ptr<void, aclError (*)(void *)> workspaceAddrPtrV3(nullptr, aclrtFree);
  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);
    workspaceAddrPtrV3.reset(workspaceAddr);
  }

  // 调用aclnnQuantMatmulV3第二段接口
  ret = aclnnQuantMatmulV3(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3 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(outShape);
  std::vector<uint16_t> resultData(size, 0); // C语言中无法直接打印fp16的数据，需要用uint16读出来，自行通过二进制转成fp16
  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("result[%ld] is: %u\n", i, resultData[i]);
  }
  return ACL_SUCCESS;
}

int main() {
  // 1. （固定写法）device/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;
  aclrtStream stream;
  auto ret = aclnnQuantMatmulV3Test(deviceId, stream);
  CHECK_FREE_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnQuantMatmulV3Test failed. ERROR: %d\n", ret); return ret);

  Finalize(deviceId, stream);
  return 0;
}