aclnnTriangularSolve
- Atlas 训练系列产品。
- Atlas A2训练系列产品/Atlas 800I A2推理产品。
接口原型
每个算子分为两段式接口,必须先调用“aclnnTriangularSolveGetWorkspaceSize”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“aclnnTriangularSolve”接口执行计算。
aclnnStatus aclnnTriangularSolveGetWorkspaceSize(const aclTensor *self, const aclTensor *A, bool upper, bool transpose, bool unitriangular, aclTensor *xOut, aclTensor *mOut, uint64_t *workspaceSize, aclOpExecutor **executor)
aclnnStatus aclnnTriangularSolve(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, const aclrtStream stream)
功能描述
- 算子功能:求解一个具有方形上或下三角形可逆矩阵A和多个右侧b的方程组。
- 计算公式: 其中是一个上三角方阵(当upper为false时为下三角方阵),其主对角线不含0的元素。为二维矩阵或者二维矩阵的batch,当输入为batch时,返回输出的X也为对应的batch。当的主对角线含有0,或元素非常接近0,且unitriangular为False时,输出结果可能包含
aclnnTriangularSolveGetWorkspaceSize
参数说明:
self(const aclTensor *, 计算输入): 公式中的,数据类型支持FLOAT、DOUBLE、COMPLEX64、COMPLEX128, 且数据类型与
A
一致,且数据维度至少为2且不大于8。支持非连续的Tensor,数据格式支持ND。A(const aclTensor *, 计算输入): 公式中的,数据类型支持FLOAT、DOUBLE、COMPLEX64、COMPLEX128, 且数据类型与
self
一致,且数据维度至少为2且不大于8。支持非连续的Tensor,数据格式支持ND。upper(bool, 计算输入):计算属性,默认为true,
A
为上三角方阵,当upper为false时,A
为下三角方阵。transpose(bool, 计算输入):计算属性,默认为false, 当transpsose为true时,计算。
unitriangular(bool, 计算输入):计算属性,默认为false,当unitriangular为true时,
A
的主对角线元素视为1,而不是从A
引用,并且unitriangular为true时输入self
和A
,输出xOut
和mOut
的数据类型只支持FLOAT。xOut(aclTensor *, 计算输出): 公式中的,数据类型支持FLOAT、DOUBLE、COMPLEX64、COMPLEX128,且数据类型与self一致,支持非连续的Tensor,数据格式支持ND,且shape需要与broadcast后的
A
,b
满足约束。mOut(aclTensor *, 计算输出): broadcast后
A
的上三角(下三角)拷贝,数据类型支持FLOAT、DOUBLE、COMPLEX64、COMPLEX128,且数据类型与self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,包含了算子计算流程。
返回值:
aclnnStatus: 返回状态码,具体参见aclnn返回码。
161001(ACLNN_ERR_PARAM_NULLPTR):1. 传入的self、A或xOut,mOut是空指针。 161002(ACLNN_ERR_PARAM_INVALID):1. self、A或xOut,mOut的数据类型和数据格式不在支持的范围之内。 2. self、A或xOut,mOut的shape不符合约束
aclnnTriangularSolve
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aclnnTriangularSolveGetWorkspaceSize获取。
executor(aclOpExecutor *, 入参):op执行器,包含了算子计算流程。
stream(aclrtStream, 入参):指定执行任务的 AscendCL Stream流。
返回值:
aclnnStatus: 返回状态码,具体参见aclnn返回码。
约束与限制
无。
调用示例
示例代码如下,仅供参考,具体编译和执行过程请参考编译与运行样例。
#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_triangular_solve.h"
#define CHECK_RET(cond, return_expr) \
do { \
if (!(cond)) { \
return_expr; \
} \
} while (0)
#define LOG_PRINT(message, ...) \
do { \
printf(message, ##__VA_ARGS__); \
} while (0)
int64_t GetShapeSize(const std::vector<int64_t>& shape) {
int64_t shapeSize = 1;
for (auto i : shape) {
shapeSize *= i;
}
return shapeSize;
}
int Init(int32_t deviceId, aclrtStream* stream) {
// 固定写法,AscendCL初始化
auto ret = aclInit(nullptr);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclInit failed. ERROR: %d\n", ret); return ret);
ret = aclrtSetDevice(deviceId);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetDevice failed. ERROR: %d\n", ret); return ret);
ret = aclrtCreateStream(stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateStream failed. ERROR: %d\n", ret); return ret);
return 0;
}
template <typename T>
int CreateAclTensor(const std::vector<T>& hostData, const std::vector<int64_t>& shape, void** deviceAddr,
aclDataType dataType, aclTensor** tensor) {
auto size = GetShapeSize(shape) * sizeof(T);
// 调用aclrtMalloc申请device侧内存
auto ret = aclrtMalloc(deviceAddr, size, ACL_MEM_MALLOC_HUGE_FIRST);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMalloc failed. ERROR: %d\n", ret); return ret);
// 调用aclrtMemcpy将host侧数据拷贝到device侧内存上
ret = aclrtMemcpy(*deviceAddr, size, hostData.data(), size, ACL_MEMCPY_HOST_TO_DEVICE);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtMemcpy failed. ERROR: %d\n", ret); return ret);
// 计算连续tensor的strides
std::vector<int64_t> strides(shape.size(), 1);
for (int64_t i = shape.size() - 2; i >= 0; i--) {
strides[i] = shape[i + 1] * strides[i + 1];
}
// 调用aclCreateTensor接口创建aclTensor
*tensor = aclCreateTensor(shape.data(), shape.size(), dataType, strides.data(), 0, aclFormat::ACL_FORMAT_ND,
shape.data(), shape.size(), *deviceAddr);
return 0;
}
int main() {
// 1. (固定写法)device/stream初始化,参考AscendCL对外接口列表
// 根据自己的实际device填写deviceId
int32_t deviceId = 0;
aclrtStream stream;
auto ret = Init(deviceId, &stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);
// 2. 构造输入与输出,需要根据API的接口自定义构造
std::vector<int64_t> selfShape = {3, 1};
std::vector<int64_t> otherShape = {3, 3};
std::vector<int64_t> xOutShape = {3, 1};
std::vector<int64_t> mOutShape = {3, 3};
void* selfDeviceAddr = nullptr;
void* otherDeviceAddr = nullptr;
void* xOutDeviceAddr = nullptr;
void* mOutDeviceAddr = nullptr;
aclTensor* self = nullptr;
aclTensor* other = nullptr;
aclTensor* xOut = nullptr;
aclTensor* mOut = nullptr;
bool upper = true;
bool transpose = false;
bool unitriangular = false;
std::vector<float> selfHostData = {1, 2, 3};
std::vector<float> otherHostData = {1, 2, 3, 0, 4, 5, 0, 0, 6};
std::vector<float> xOutHostData = {-0.2500, -0.1250, 0.5000};
std::vector<float> mOutHostData = {1, 2, 3, 0, 4, 5, 0, 0, 6};
// 创建self aclTensor
ret = CreateAclTensor(selfHostData, selfShape, &selfDeviceAddr, aclDataType::ACL_FLOAT, &self);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建other aclTensor
ret = CreateAclTensor(otherHostData, otherShape, &otherDeviceAddr, aclDataType::ACL_FLOAT, &other);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建xOut aclTensor
ret = CreateAclTensor(xOutHostData, xOutShape, &xOutDeviceAddr, aclDataType::ACL_FLOAT, &xOut);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 创建mOut aclTensor
ret = CreateAclTensor(mOutHostData, mOutShape, &mOutDeviceAddr, aclDataType::ACL_FLOAT, &mOut);
CHECK_RET(ret == ACL_SUCCESS, return ret);
// 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aclnnTriangularSolve第一段接口
ret = aclnnTriangularSolveGetWorkspaceSize(self, other, upper, transpose, unitriangular, xOut, mOut, &workspaceSize, &executor);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTriangularSolveGetWorkspaceSize 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aclnnTriangularSolve第二段接口
ret = aclnnTriangularSolve(workspaceAddr, workspaceSize, executor, stream);
CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnTriangularSolve 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 xSize = GetShapeSize(xOutShape);
std::vector<float> resultData(xSize, 0);
ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]), xOutDeviceAddr,
xSize * 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 < xSize; i++) {
LOG_PRINT("result[%ld] is: %f\n", i, resultData[i]);
}
auto mSize = GetShapeSize(mOutShape);
std::vector<float> mResultData(mSize, 0);
ret = aclrtMemcpy(mResultData.data(), mResultData.size() * sizeof(mResultData[0]), mOutDeviceAddr,
mSize * sizeof(mResultData[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 < mSize; i++) {
LOG_PRINT("result[%ld] is: %f\n", i, mResultData[i]);
}
// 6. 释放aclTensor和aclScalar,需要根据具体API的接口定义修改
aclDestroyTensor(self);
aclDestroyTensor(other);
aclDestroyTensor(xOut);
aclDestroyTensor(mOut);
// 7. 释放device资源,需要根据具体API的接口定义修改
aclrtFree(selfDeviceAddr);
aclrtFree(otherDeviceAddr);
aclrtFree(xOutDeviceAddr);
aclrtFree(mOutDeviceAddr);
if (workspaceSize > 0) {
aclrtFree(workspaceAddr);
}
aclrtDestroyStream(stream);
aclrtResetDevice(deviceId);
aclFinalize();
return 0;
}