文档HOT
下载
中文
注册

aclnnApplyRotaryPosEmb

支持的产品型号

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

接口原型

每个算子分为两段式接口，必须先调用“aclnnApplyRotaryPosEmbGetWorkspaceSize”接口获取入参并根据流程计算所需workspace大小，再调用“aclnnApplyRotaryPosEmb”接口执行计算。

  • aclnnStatus aclnnApplyRotaryPosEmbGetWorkspaceSize(aclTensor *queryRef, aclTensor *keyRef, const aclTensor *cos, const aclTensor *sin, int64_t layoutOptional, uint64_t *workspaceSize, aclOpExecutor **executor)
  • aclnnStatus aclnnApplyRotaryPosEmb(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor， aclrtStream stream)

功能描述

  • 算子功能：推理网络为了提升性能，将query和key两路算子融合成一路。
  • 计算公式：
query_q1=query[...,:query.shape[−1]//2]query\_q1 = query[..., : query.shape[-1] // 2] query_q2=query[...,query.shape[−1]//2:]query\_q2 = query[..., query.shape[-1] // 2 :] query_rotate=torch.cat((−query_q2,query_q1),dim=−1)query\_rotate = torch.cat((-query\_q2, query\_q1), dim=-1) key_k1=key[...,:key.shape[−1]//2]key\_k1 = key[..., : key.shape[-1] // 2] key_k2=key[...,key.shape[−1]//2:]key\_k2 = key[..., key.shape[-1] // 2 :] key_rotate=torch.cat((−key_k2,key_k1),dim=−1)key\_rotate = torch.cat((-key\_k2, key\_k1), dim=-1) q_embed=(query∗cos)+query_rotate∗sinq\_embed = (query * cos) + query\_rotate * sin k_embed=(key∗cos)+key_rotate∗sink\_embed = (key * cos) + key\_rotate * sin

aclnnApplyRotaryPosEmbGetWorkspaceSize

  • 参数说明：

    • queryRef(aclTensor*, 计算输入)：输入Tensor，数据类型支持BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）、FLOAT16、FLOAT32。支持非连续的Tensor，数据格式支持ND。
    • keyRef(aclTensor*, 计算输入)：输入Tensor，数据类型支持BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）、FLOAT16、FLOAT32。支持非连续的Tensor，数据格式支持ND。
    • cos(const aclTensor*, 计算输入)：输入Tensor，数据类型支持BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）、FLOAT16、FLOAT32。支持非连续的Tensor，数据格式支持ND。
    • sin(const aclTensor*, 计算输入)：输入Tensor，数据类型支持BFLOAT16（仅Atlas A2训练系列产品/Atlas 800I A2推理产品支持）、FLOAT16、FLOAT32。支持非连续的Tensor，数据格式支持ND。
    • layoutOptional(int64_t, 计算输入)：公式中的布局格式，数据类型int64, 目前只支持1, 代表"BSH"格式，实际代表支持格式为BSND，4维。
    • 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。

  • 说明： 第一段接口完成入参校验，若出现以下错误码，则对应原因为：

    • 返回161001（ACLNN_ERR_PARAM_NULLPTR）：如果传入参数是必选输入，输出或者必选属性，且是空指针，则返回161001。
    • 返回161002（ACLNN_ERR_PARAM_INVALID）：queryRef、keyRef、cos、sin、layoutOptional的数据类型和数据格式不在支持的范围内。

aclnnApplyRotaryPosEmb

  • 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aclnnApplyRotaryPosEmbGetWorkspaceSize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

  • 输入张量queryRef、keyRef、cos、sin只支持4维的shape，layoutOptional只支持1
  • 输入张量queryRef、keyRef、cos、sin 的dtype必须相同，且4个输入shape的前2维和最后一维必须相等，cos和sin的shape第3维必须等于1，输入shape最后一维必须等于128
  • 输入queryRef的shape用（q_b,q_s,q_n,q_d）表示，keyRef shape用（q_b,q_s,k_n,q_d）,cos和sin shape用（q_b,q_s,1,q_d），当输入是BFLOAT16时，cast表示为1，castSize为4， DtypeSize为2，不是BFLOAT16时，cast表示为0，castSize = DtypeSize，float16时为2，float32时为4， 需要使用的ub大小为：ub = (q_n + k_n)128 castSize2 + 128DtypeSize4 + (q_n + k_n)128castSize + (q_n + k_n)128castSize2 + cast*(12842) 当ub大于对应可以获取到的AI处理器版本总ub大小时，不支持。
  • 不支持空tensor场景

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。

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

#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() {
    // 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 = Init(deviceId, &stream);
    // check根据自己的需要处理
    CHECK_RET(ret == 0, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);
    // 2. 构造输入与输出，需要根据API的接口定义构造
    std::vector<int64_t> queryShape = {1, 1, 1, 128};
    std::vector<int64_t> keyShape = {1, 1, 1, 128};
    std::vector<int64_t> cosShape = {1, 1, 1, 128};
    std::vector<int64_t> sinShape = {1, 1, 1, 128};
    int64_t layout = 1;

    void* queryDeviceAddr = nullptr;
    void* keyDeviceAddr = nullptr;
    void* cosDeviceAddr = nullptr;
    void* sinDeviceAddr = nullptr;
    aclTensor* query = nullptr;
    aclTensor* key = nullptr;
    aclTensor* cos = nullptr;
    aclTensor* sin = nullptr;

    std::vector<float> queryHostData = {74, 54, 84, 125, 23, 78, 37, 72, 27, 98, 34, 107, 29, 23, 54, 60, 70, 49,
                                        119, 54, 29, 54, 41, 99, 27, 62, 5, 46, 108, 39, 24, 123, 33, 82, 6, 40, 88,
                                        24, 6, 116, 38, 119, 110, 5, 30, 79, 87, 18, 29, 100, 90, 24, 21, 93, 63, 68,
                                        34, 112, 119, 48, 74, 43, 85, 64, 14, 49, 128, 59, 18, 37, 123, 76, 14, 63, 10,
                                        39, 107, 124, 79, 16, 17, 76, 80, 47, 90, 41, 58, 82, 75, 80, 69, 37, 74, 36, 54,
                                        26, 32, 54, 13, 100, 105, 15, 13, 69, 122, 26, 94, 59, 29, 14, 60, 8, 24, 17, 45,
                                        33, 107, 122, 63, 111, 75, 128, 68, 31, 105, 6, 82, 99};
    std::vector<float> keyHostData = {112, 32, 66, 114, 69, 31, 117, 122, 77, 57, 78, 119, 115, 25, 54, 27, 122, 65, 15, 85,
                                      33, 16, 36, 6, 95, 15, 43, 6, 66, 91, 14, 101, 78, 51, 110, 74, 56, 30, 127, 61, 53, 29,
                                      32, 65, 114, 77, 26, 116, 89, 38, 75, 14, 96, 91, 87, 34, 25, 42, 57, 26, 51, 43, 23, 42,
                                      40, 17, 98, 117, 53, 75, 68, 75, 38, 41, 115, 76, 67, 22, 76, 10, 24, 46, 85, 54, 61, 114,
                                      10, 59, 6, 123, 58, 10, 115, 9, 13, 58, 66, 120, 23, 30, 83, 13, 11, 76, 18, 82, 57, 4,
                                      117, 105, 8, 73, 127, 5, 91, 56, 12, 125, 20, 3, 104, 40, 46, 18, 89, 63, 99, 104};
    std::vector<float> cosHostData = {41, 37, 17, 25, 49, 25, 22, 24, 110, 120, 107, 3, 82, 66, 75, 86, 85, 115, 110, 56, 52,
                                      39, 86, 23, 36, 71, 20, 73, 113, 25, 114, 56, 125, 80, 95, 82, 31, 63, 99, 62, 23, 55, 30,
                                      99, 42, 121, 15, 24, 97, 87, 81, 67, 43, 21, 13, 9, 33, 29, 117, 10, 114, 61, 98, 15, 78,
                                      108, 48, 97, 1, 3, 78, 109, 57, 46, 47, 56, 50, 66, 81, 77, 17, 128, 68, 121, 47, 91, 114,
                                      125, 51, 108, 31, 15, 47, 78, 109, 115, 113, 26, 53, 97, 1, 111, 103, 58, 106, 68, 11,
                                      104, 22, 79, 61, 127, 86, 39, 33, 123, 102, 39, 64, 41, 119, 120, 61, 29, 94, 68, 36, 12};
    std::vector<float> sinHostData = {46, 56, 56, 101, 66, 10, 96, 16, 86, 57, 102, 66, 12, 105, 76, 58, 90, 6, 79, 128, 126,
                                      82, 41, 3, 45, 7, 66, 4, 46, 22, 31, 26, 37, 63, 97, 84, 91, 90, 47, 77, 90, 34, 41, 83,
                                      91, 108, 120, 13, 90, 32, 85, 37, 119, 31, 51, 82, 122, 125, 7, 116, 121, 108, 38, 56,
                                      100, 20, 97, 119, 10, 4, 53, 13, 46, 82, 103, 119, 124, 80, 23, 67, 78, 56, 119, 122, 40,
                                      58, 128, 27, 30, 52, 71, 42, 123, 69, 4, 5, 116, 97, 38, 107, 8, 4, 65, 120, 40, 22, 60,
                                      44, 48, 66, 68, 125, 4, 93, 112, 112, 113, 90, 94, 23, 104, 39, 85, 84, 64, 128, 96, 119};

    // 创建query aclTensor
    ret = CreateAclTensor(queryHostData, queryShape, &queryDeviceAddr, aclDataType::ACL_FLOAT, &query);
    CHECK_RET(ret == ACL_SUCCESS, return ret);
    // 创建key aclTensor
    ret = CreateAclTensor(keyHostData, keyShape, &keyDeviceAddr, aclDataType::ACL_FLOAT, &key);
    CHECK_RET(ret == ACL_SUCCESS, return ret);
    // 创建cos aclTensor
    ret = CreateAclTensor(cosHostData, cosShape, &cosDeviceAddr, aclDataType::ACL_FLOAT, &cos);
    CHECK_RET(ret == ACL_SUCCESS, return ret);
    // 创建sin aclTensor
    ret = CreateAclTensor(sinHostData, sinShape, &sinDeviceAddr, aclDataType::ACL_FLOAT, &sin);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // 3. 调用CANN算子库API，需要修改为具体的API
    uint64_t workspaceSize = 0;
    aclOpExecutor* executor;
    // 调用aclnnApplyRotaryPosEmb第一段接口
    ret = aclnnApplyRotaryPosEmbGetWorkspaceSize(query, key, cos, sin, layout, &workspaceSize, &executor);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnApplyRotaryPosEmbGetWorkspaceSize 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aclnnApplyRotaryPosEmb第二段接口
    ret = aclnnApplyRotaryPosEmb(workspaceAddr, workspaceSize, executor, stream);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnApplyRotaryPosEmb 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(queryShape);
    std::vector<float> resultData(size, 0);
    ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), queryDeviceAddr, size * sizeof(float),
                      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: %hhu\n", i, resultData[i]);
    }

    auto size1 = GetShapeSize(keyShape);
    std::vector<float> resultData1(size, 0);
    ret = aclrtMemcpy(resultData1.data(), resultData1.size() * sizeof(resultData1[0]), keyDeviceAddr, size1 * sizeof(float),
                      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: %hhu\n", i, resultData[i]);
    }

    // 6. 释放aclTensor和aclScalar，需要根据具体API的接口定义修改
    aclDestroyTensor(query);
    aclDestroyTensor(key);
    aclDestroyTensor(cos);
    aclDestroyTensor(sin);

    // 7. 释放device 资源
    aclrtFree(queryDeviceAddr);
    aclrtFree(keyDeviceAddr);
    aclrtFree(cosDeviceAddr);
    aclrtFree(sinDeviceAddr);
    if (workspaceSize > 0) {
      aclrtFree(workspaceAddr);
    }
    aclrtDestroyStream(stream);
    aclrtResetDevice(deviceId);
    aclFinalize();

    return 0;
}