aclnnRingAttentionUpdate
支持的产品型号
- 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“aclnnRingAttentionUpdateGetWorkspaceSize”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“aclnnRingAttentionUpdate”接口执行计算。
aclnnStatus aclnnRingAttentionUpdateGetWorkspaceSize(const aclTensor* prevAttnOut, const aclTensor* prevSoftmaxMax, const aclTensor* prevSoftmaxSum, const aclTensor* curAttnOut, const aclTensor* curSoftmaxMax, const aclTensor* curSoftmaxSum, const aclTensor* actualSeqQlenOptional, char* inputLayout, aclTensor* attnOut, aclTensor* softmaxMax, aclTensor* softmaxSum, uint64_t* workspaceSize, aclOpExecutor** executor)
aclnnStatus aclnnRingAttentionUpdate(void* workspace, uint64_t workspaceSize, aclOpExecutor* executor, aclrtStream stream)
功能描述
接口功能:RingAttentionUpdate算子功能是将两次FlashAttention的输出根据其不同的softmax的max和sum更新。
计算公式:
aclnnRingAttentionUpdateGetWorkspaceSize
参数说明:
- prevAttnOut(aclTensor*,计算输入):Device侧的aclTensor,公式中的prev_attn_out,数据类型支持FLOAT16、FLOAT、BFLOAT16,输入shape和inputLayout属性保持一致,支持非连续的Tensor,数据格式支持ND。
- prevSoftmaxMax(aclTensor*,计算输入):Device侧的aclTensor,公式中的prev_softmax_max,数据类型支持FLOAT,输入shape为(B,N,S,8),最后一维8个数字相同,且需要为正数,支持非连续的Tensor,数据格式支持ND。
- prevSoftmaxSum(aclTensor*,计算输入):Device侧的aclTensor,公式中的prev_softmax_sum,数据类型支持FLOAT,输入shape和prevSoftmaxMax保持一致,最后一维8个数字相同,且需要为正数,支持非连续的Tensor,数据格式支持ND。
- curAttnOut(aclTensor*,计算输入):Device侧的aclTensor,公式中的cur_attn_out,数据类型支持FLOAT16、FLOAT、BFLOAT16,数据类型和输入shape和prevAttnOut保持一致,支持非连续的Tensor,数据格式支持ND。
- curSoftmaxMax(aclTensor*,计算输入):Device侧的aclTensor,公式中的cur_softmax_max,数据类型支持FLOAT,输入shape和prevSoftmaxMax保持一致,最后一维8个数字相同,且需要为正数,支持非连续的Tensor,数据格式支持ND。
- curSoftmaxSum(aclTensor*,计算输入):Device侧的aclTensor,公式中的cur_softmax_sum,数据类型支持FLOAT,输入shape和prevSoftmaxMax保持一致,最后一维8个数字相同,且需要为正数,支持非连续的Tensor,数据格式支持ND。
- actualSeqQlenOptional(aclTensor*,计算输入):Device侧的aclTensor,预留接口,暂时无效,当有输入时,数据类型支持INT64。
- inputLayout(char*,计算输入):Host侧的char*常量,attn_out相关输入的数据排布。预留接口,暂时无效。
- attnOut(aclTensor*,计算输出):Device侧的aclTensor,公式中的attn_out,数据类型支持FLOAT16、FLOAT、BFLOAT16,数据类型和输出shape和prevAttnOut保持一致,支持非连续的Tensor,数据格式支持ND。
- softmaxMax(aclTensor*,计算输出):Device侧的aclTensor,公式中的softmax_max,数据类型支持FLOAT,输出shape和prevSoftmaxMax保持一致,支持非连续的Tensor,数据格式支持ND。
- softmaxSum(aclTensor*,计算输出):Device侧的aclTensor,公式中的softmax_sum,数据类型支持FLOAT,输出shape和prevSoftmaxMax保持一致,支持非连续的Tensor,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: 返回161001(ACLNN_ERR_PARAM_NULLPTR):1. 传入的 prevAttnOut、prevSoftmaxMax、prevSoftmaxSum、curAttnOut、curSoftmaxMax、curSoftmaxSum、attnOut、softmaxMax、softmaxSum 是空指针时。 返回161002(ACLNN_ERR_PARAM_INVALID):1. prevAttnOut、prevSoftmaxMax、prevSoftmaxSum、curAttnOut、curSoftmaxMax、curSoftmaxSum、attnOut、softmaxMax、softmaxSum数据类型不在支持的范围之内。 2. prevAttnOut、prevSoftmaxMax、prevSoftmaxSum、curAttnOut、curSoftmaxMax、curSoftmaxSum、attnOut、softmaxMax、softmaxSum的shape不支持。 返回561002 (ACLNN_ERR_INNER_TILING_ERROR):1. 当actualSeqQlenOptional有输入时,输入数据格式不在支持的范围之内
aclnnRingAttentionUpdate
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aclnnRingAttentionUpdateGetWorkspaceSize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
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 <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_ring_attention_update.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() {
// 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的接口自定义构造
int64_t batchNum = 1;
int64_t headNum = 1;
int64_t seqSize = 2;
int64_t headDim = 4;
int64_t headSize = headNum * headDim;
std::vector<int64_t> prevAttnOutShape = {seqSize, batchNum, headSize};
std::vector<int64_t> prevSoftmaxMaxShape = {batchNum, headNum, seqSize, 8};
std::vector<int64_t> prevSoftmaxSumShape = {batchNum, headNum, seqSize, 8};
std::vector<int64_t> curAttnOutShape = {seqSize, batchNum, headSize};
std::vector<int64_t> curSoftmaxMaxShape = {batchNum, headNum, seqSize, 8};
std::vector<int64_t> curSoftmaxSumShape = {batchNum, headNum, seqSize, 8};
std::vector<int64_t> actualSeqQlenOptionalShape = {batchNum, headNum};
std::vector<int64_t> attnOutShape = {seqSize, batchNum, headSize};
std::vector<int64_t> softmaxMaxShape = {batchNum, headNum, seqSize, 8};
std::vector<int64_t> softmaxSumShape = {batchNum, headNum, seqSize, 8};
void* prevAttnOutDeviceAddr = nullptr;
void* prevSoftmaxMaxDeviceAddr = nullptr;
void* prevSoftmaxSumDeviceAddr = nullptr;
void* curAttnOutDeviceAddr = nullptr;
void* curSoftmaxMaxDeviceAddr = nullptr;
void* curSoftmaxSumDeviceAddr = nullptr;
void* actualSeqQlenOptionalDeviceAddr = nullptr;
void* attnOutDeviceAddr = nullptr;
void* softmaxMaxDeviceAddr = nullptr;
void* softmaxSumDeviceAddr = nullptr;
aclTensor* prevAttnOut = nullptr;
aclTensor* prevSoftmaxMax = nullptr;
aclTensor* prevSoftmaxSum = nullptr;
aclTensor* curAttnOut = nullptr;
aclTensor* curSoftmaxMax = nullptr;
aclTensor* curSoftmaxSum = nullptr;
aclTensor* actualSeqQlenOptional = nullptr;
aclTensor* attnOut = nullptr;
aclTensor* softmaxMax = nullptr;
aclTensor* softmaxSum = nullptr;
std::vector<float> prevAttnOutHostData(seqSize * batchNum * headSize, 1);
std::vector<float> prevSoftmaxMaxHostData(batchNum * headNum * seqSize * 8, 1);
std::vector<float> prevSoftmaxSumHostData(batchNum * headNum * seqSize * 8, 1);
std::vector<float> curAttnOutHostData(seqSize * batchNum * headSize, 1);
std::vector<float> curSoftmaxMaxHostData(batchNum * headNum * seqSize * 8, 1);
std::vector<float> curSoftmaxSumHostData(batchNum * headNum * seqSize * 8, 1);
std::vector<float> actualSeqQlenOptionalHostData(batchNum * headNum, 1);
std::vector<float> attnOutHostData(seqSize * batchNum * headSize, 1);
std::vector<float> softmaxMaxHostData(batchNum * headNum * seqSize * 8, 1);
std::vector<float> softmaxSumHostData(batchNum * headNum * seqSize * 8, 1);
char* inputLayout = "SBH";
// 创建prevAttnOut aclTensor
ret = CreateAclTensor(prevAttnOutHostData, prevAttnOutShape, &prevAttnOutDeviceAddr, aclDataType::ACL_FLOAT, &prevAttnOut);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建prevSoftmaxMax aclTensor
ret = CreateAclTensor(prevSoftmaxMaxHostData, prevSoftmaxMaxShape, &prevSoftmaxMaxDeviceAddr, aclDataType::ACL_FLOAT, &prevSoftmaxMax);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建prevSoftmaxSum aclTensor
ret = CreateAclTensor(prevSoftmaxSumHostData, prevSoftmaxSumShape, &prevSoftmaxSumDeviceAddr, aclDataType::ACL_FLOAT, &prevSoftmaxSum);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建curAttnOut aclTensor
ret = CreateAclTensor(curAttnOutHostData, curAttnOutShape, &curAttnOutDeviceAddr, aclDataType::ACL_FLOAT, &curAttnOut);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建curSoftmaxMax aclTensor
ret = CreateAclTensor(curSoftmaxMaxHostData, curSoftmaxMaxShape, &curSoftmaxMaxDeviceAddr, aclDataType::ACL_FLOAT, &curSoftmaxMax);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建curSoftmaxSum aclTensor
ret = CreateAclTensor(curSoftmaxSumHostData, curSoftmaxSumShape, &curSoftmaxSumDeviceAddr, aclDataType::ACL_FLOAT, &curSoftmaxSum);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建actualSeqQlenOptional aclTensor
ret = CreateAclTensor(actualSeqQlenOptionalHostData, actualSeqQlenOptionalShape, &actualSeqQlenOptionalDeviceAddr, aclDataType::ACL_INT64, &actualSeqQlenOptional);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建attnOut aclTensor
ret = CreateAclTensor(attnOutHostData, attnOutShape, &attnOutDeviceAddr, aclDataType::ACL_FLOAT, &attnOut);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建softmaxMax aclTensor
ret = CreateAclTensor(softmaxMaxHostData, softmaxMaxShape, &softmaxMaxDeviceAddr, aclDataType::ACL_FLOAT, &softmaxMax);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建softmaxSum aclTensor
ret = CreateAclTensor(softmaxSumHostData, softmaxSumShape, &softmaxSumDeviceAddr, aclDataType::ACL_FLOAT, &softmaxSum);
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aclnnRingAttentionUpdate第一段接口
ret = aclnnRingAttentionUpdateGetWorkspaceSize(prevAttnOut, prevSoftmaxMax, prevSoftmaxSum,
curAttnOut, curSoftmaxMax, curSoftmaxSum,
actualSeqQlenOptional, inputLayout,
attnOut, softmaxMax, softmaxSum, &workspaceSize, &executor);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnRingAttentionUpdateGetWorkspaceSize 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aclnnRingAttentionUpdate第二段接口
ret = aclnnRingAttentionUpdate(workspaceAddr, workspaceSize, executor, stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnRingAttentionUpdate 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 attnOutSize = GetShapeSize(attnOutShape);
std::vector<float> attnOutResultData(attnOutSize, 0);
ret = aclrtMemcpy(attnOutResultData.data(), attnOutResultData.size() * sizeof(attnOutResultData[0]), attnOutDeviceAddr, attnOutSize * 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 < attnOutSize; i++) {
LOG_PRINT("attnOutResultData[%ld] is: %f\n", i, attnOutResultData[i]);
}
auto softmaxMaxSize = GetShapeSize(softmaxMaxShape);
std::vector<float> softmaxMaxResultData(softmaxMaxSize, 0);
ret = aclrtMemcpy(softmaxMaxResultData.data(), softmaxMaxResultData.size() * sizeof(softmaxMaxResultData[0]), softmaxMaxDeviceAddr, softmaxMaxSize * 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 < softmaxMaxSize; i++) {
LOG_PRINT("softmaxMaxResultData[%ld] is: %f\n", i, softmaxMaxResultData[i]);
}
auto softmaxSumSize = GetShapeSize(softmaxSumShape);
std::vector<float> softmaxSumResultData(softmaxSumSize, 0);
ret = aclrtMemcpy(softmaxSumResultData.data(), softmaxSumResultData.size() * sizeof(softmaxSumResultData[0]), softmaxSumDeviceAddr, softmaxSumSize * 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 < softmaxSumSize; i++) {
LOG_PRINT("softmaxSumResultData[%ld] is: %f\n", i, softmaxSumResultData[i]);
}
// 6. 释放aclTensor和aclScalar,需要根据具体API的接口定义修改
aclDestroyTensor(prevAttnOut);
aclDestroyTensor(prevSoftmaxMax);
aclDestroyTensor(prevSoftmaxSum);
aclDestroyTensor(curAttnOut);
aclDestroyTensor(curSoftmaxMax);
aclDestroyTensor(curSoftmaxSum);
aclDestroyTensor(actualSeqQlenOptional);
aclDestroyTensor(attnOut);
aclDestroyTensor(softmaxMax);
aclDestroyTensor(softmaxSum);
// 7. 释放device资源,需要根据具体API的接口定义修改
aclrtFree(prevAttnOutDeviceAddr);
aclrtFree(prevSoftmaxMaxDeviceAddr);
aclrtFree(prevSoftmaxSumDeviceAddr);
aclrtFree(curAttnOutDeviceAddr);
aclrtFree(curSoftmaxMaxDeviceAddr);
aclrtFree(curSoftmaxSumDeviceAddr);
aclrtFree(actualSeqQlenOptionalDeviceAddr);
aclrtFree(attnOutDeviceAddr);
aclrtFree(softmaxMaxDeviceAddr);
aclrtFree(softmaxSumDeviceAddr);
if (workspaceSize > 0) {
aclrtFree(workspaceAddr);
}
aclrtDestroyStream(stream);
aclrtResetDevice(deviceId);
aclFinalize();
return 0;
}