aclnnCdistBackward

Applicable Products

Product	Supported
Ascend 950PR/Ascend 950DT	×
Atlas A3 training products/Atlas A3 inference products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference products	×
Atlas inference products	×
Atlas training products	×

Function

Description: Completes the backward propagation of aclnnCdist.
Formula:
$\begin{aligned} out&=grad \cdot y' \\ &= grad \cdot \left( \sqrt[p]{\sum (x_1 - x_2)^p} \right)' \\ &= grad \cdot \frac{1}{p} \times \left( \sum (x_1 - x_2)^p \right)^{\frac{1}{p}-1} \times p \times (x_1 - x_2)^{p-1} \\ &= grad \cdot \left( \sum (x_1 - x_2)^p \right)^{\frac{-(p-1)}{p}} \times (x_1 - x_2)^{p-1} \\ &= grad \cdot \left( \sum (x_1 - x_2)^p \right)^{\frac{1}{p} \times (-(p-1))} \times (x_1 - x_2)^{p-1} \\ &= grad \cdot \frac{diff^{p-1}}{cdist^{p-1}} \\ &= grad \cdot \frac{diff \times |diff|^{p-2}}{cdist^{p-1}} \end{aligned}$
- $\mathrm{diff} = x_1 - x_2$ : variable difference
- $\mathrm{cdist} = \sqrt[p]{\sum (x_1 - x_2)^p}$ : $p$ -norm distance

Function Prototype

Each operator is divided into a two-phase API. You must first call the "aclnnCdistBackwardGetWorkspaceSize" API to obtain the input parameters and calculate the required workspace size based on the process, and then call the "aclnnCdistBackward" API to perform the computation.

aclnnStatus aclnnCdistBackwardGetWorkspaceSize(
    const aclTensor *grad,
    const aclTensor *x1,
    const aclTensor *x2,
    const aclTensor *cdist,
    float            p,
    aclTensor       *out,
    uint64_t        *workspaceSize,
    aclOpExecutor  **executor)

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

aclnnCdistBackwardGetWorkspaceSize

Parameters:

Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (shape)	Non-contiguous Tensor
grad (aclTensor*)	Input	grad in the formula.	-	FLOAT, FLOAT16, BFLOAT16	ND	2-7 dimensions.	√
x1（aclTensor*）	Input	x1 in the formula.	Data type is consistent with grad.	FLOAT, FLOAT16, BFLOAT16	ND	Has the same number of dimensions as grad, and the shape is consistent with grad except for the last dimension.	√
x2 (aclTensor*)	Input	x2 in the formula.	The data type is consistent with grad.	FLOAT, FLOAT16, BFLOAT16	ND	The number of dimensions equals that of grad. The second-to-last dimension equals the last dimension of grad. The last dimension equals the last dimension of x1. The remaining dimensions must satisfy the broadcast relationship with grad.	√
cdist (aclTensor*)	Input	cdist in the formula.	Data type consistent with grad.	FLOAT, FLOAT16, BFLOAT16	ND	Shape consistent with grad.	√
p (float)	Attribute	p in the formula.	-	float	-	-	-
out (aclTensor*)	Output	out in the formula.	The data type is consistent with grad.	FLOAT, FLOAT16, BFLOAT16	ND	Shape is consistent with x1.	√
workspaceSize (uint64_t*)	Output	Size of the workspace to be applied on the Device side.	-	-	-	-	-
executor (aclOpExecutor**)	Output	Returns the op executor, which contains the operator computation flow.	-	-	-	-	-

Return value

aclnnStatus: return code. For details, see aclnn Return Codes.

The first-phase API performs input parameter validation and returns an error in the following scenarios:

Return Code	Error Code	Description
ACLNN_ERR_PARAM_NULLPTR	161001	Indicates that grad, x1, x2, or cdist is a null pointer.
ACLNN_ERR_PARAM_INVALID	161002	The data type of grad, x1, x2, or cdist is not within the supported range.
		The shape of grad, x1, x2, or cdist is not within the supported range.

aclnnCdistBackward

Parameters:

Parameter	Input/Output	Description
workspace	Input	Address of the workspace memory applied on the Device side.
workspaceSize	Input	Size of the workspace applied on the Device side, obtained through the first-phase API aclnnCdistBackwardGetWorkspaceSize.
executor	Input	Operator executor that contains the operator computation flow.
stream	Input	Specifies the stream for task execution.

Return value

aclnnStatus: return code. For details, see aclnn Return Codes.

Constraints

Defaults to a deterministic computation.

Example

The following provides example code for reference only. For details about compilation and running, see Compile and Run Samples.

#include <iostream>
#include <vector>
#include "acl/acl.h"
#include "aclnnop/aclnn_cdist_backward.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)
{
    // Boilerplate: resource initialization.
    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);
    // Call aclrtMalloc to allocate memory on the device side.
    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);
    // Call aclrtMemcpy to copy data from the host side to the device memory.
    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);

    // Compute the strides of a contiguous tensor.
    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];
    }

    // Call the aclCreateTensor API to create an 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. (Boilerplate) Initialize the device/stream. See the ACL API manual.
    // Set deviceId based on the actual device.
    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);
    int B=5, P=7, Q=9, M=11;
    // 2. Construct the input and output. Customize the construction based on the API.
    std::vector<int64_t> gradShape = {B, P, Q};
    std::vector<int64_t> x1Shape = {B, P, M};
    std::vector<int64_t> x2Shape = {B, Q, M};
    std::vector<int64_t> cdistShape = {B, P, Q};
    std::vector<int64_t> outShape = {B, P, M};
    void *gradDeviceAddr = nullptr;
    void *x1DeviceAddr = nullptr;
    void *x2DeviceAddr = nullptr;
    void *cdistDeviceAddr = nullptr;
    void *outDeviceAddr = nullptr;
    aclTensor *grad = nullptr;
    aclTensor *x1 = nullptr;
    aclTensor *x2 = nullptr;
    aclTensor *cdist = nullptr;
    aclTensor *out = nullptr;
    float p = 0.5;
    std::vector<float> gradHostData(B * P * Q, 1);
    std::vector<float> x1HostData(B * P * M, 2);
    std::vector<float> x2HostData(B * Q * M, 4);
    std::vector<float> cdistHostData(B * P * Q, 0.5);
    std::vector<float> outHostData(B * P * M, 1);
    // Create the grad aclTensor.
    ret = CreateAclTensor(gradHostData, gradShape, &gradDeviceAddr, aclDataType::ACL_FLOAT, &grad);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // Create the x1 aclTensor.
    ret = CreateAclTensor(x1HostData, x1Shape, &x1DeviceAddr, aclDataType::ACL_FLOAT, &x1);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // Create the x2 aclTensor.
    ret = CreateAclTensor(x2HostData, x2Shape, &x2DeviceAddr, aclDataType::ACL_FLOAT, &x2);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // Create the cdist aclTensor.
    ret = CreateAclTensor(cdistHostData, cdistShape, &cdistDeviceAddr, aclDataType::ACL_FLOAT, &cdist);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // Create the out aclTensor.
    ret = CreateAclTensor(outHostData, outShape, &outDeviceAddr, aclDataType::ACL_FLOAT, &out);
    CHECK_RET(ret == ACL_SUCCESS, return ret);

    // 3. Call the CANN operator library API. Replace with the actual API name.
    uint64_t workspaceSize = 0;
    aclOpExecutor *executor;
    // Call the first-phase API of aclnnCdistBackward.
    ret = aclnnCdistBackwardGetWorkspaceSize(grad, x1, x2, cdist, p, out, &workspaceSize, &executor);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnCdistBackwardGetWorkspaceSize failed. ERROR: %d\n", ret); return ret);
    // Apply for device memory based on the workspaceSize calculated by the first-phase API.
    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);
    }
    // Call the second-phase API of aclnnCdistBackward.
    ret = aclnnCdistBackward(workspaceAddr, workspaceSize, executor, stream);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnCdistBackward failed. ERROR: %d\n", ret); return ret);

    // 4. (Boilerplate) Synchronously wait for task execution to complete.
    ret = aclrtSynchronizeStream(stream);
    CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSynchronizeStream failed. ERROR: %d\n", ret); return ret);

    // 5. Obtain the output value and copy the result from the device memory to the host memory. Modify based on the specific API definition.
    auto size = GetShapeSize(outShape);
    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 < 10; i++) {
        LOG_PRINT("result[%ld] is: %f\n", i, resultData[i]);
    }

    // 6. Release aclTensor and aclScalar. Modify based on the specific API definition.
    aclDestroyTensor(x1);
    aclDestroyTensor(x2);
    aclDestroyTensor(cdist);
    aclDestroyTensor(grad);
    aclDestroyTensor(out);

    // 7. Release device resources.
    aclrtFree(x1DeviceAddr);
    aclrtFree(x2DeviceAddr);
    aclrtFree(cdistDeviceAddr);
    aclrtFree(gradDeviceAddr);
    aclrtFree(outDeviceAddr);
    if (workspaceSize > 0) {
        aclrtFree(workspaceAddr);
    }
    aclrtDestroyStream(stream);
    aclrtResetDevice(deviceId);
    aclFinalize();

    return 0;
}