aclnnFlashAttentionUnpaddingScoreGradV5

Applicable Products

Function

• Description: Implements self-attention computation using the FlashAttention algorithm in training scenarios. Adds the sinkInOptional optional input.

  • Ascend 950PR/Ascend 950DT: The sinkInOptional parameter is not currently supported.

• Formula:

  The forward attention formula is:

  $$=Dropout(Softmax(Mask(\frac{QK^T+pse}{\sqrt{d}}),atten\_mask),keep\_prob)V$$

  For ease of expression, variables $S$ and $P$ are used to represent the formulas:

  $$=Mask(\frac{QK^T+pse}{\sqrt{d}}),atten\_mask$$ $$=Dropout(Softmax(S),keep\_prob)$$ $$=PV$$

  The backward propagation formula for attention is:

  $$V=P^TdY$$ $$Q=\frac{((dS)*K)}{\sqrt{d}}$$ $$K=\frac{((dS)^T*Q)}{\sqrt{d}}$$

  The computation logic after adding sink is shown below, which mainly modifies the softmax_max and softmax_sum logic computation parts.

  $$S = Q @ K^{T}$$ $$m = max(sink, max(S))$$

Function Prototype

Each operator is divided into a two-phase API. You must first call the "aclnnFlashAttentionUnpaddingScoreGradV5GetWorkspaceSize" API to obtain the workspace size required for computation and the executor that contains the operator computation process, and then call the "aclnnFlashAttentionUnpaddingScoreGradV5" API to perform the computation.

aclnnStatus aclnnFlashAttentionUnpaddingScoreGradV5GetWorkspaceSize(
  const aclTensor   *query,
  const aclTensor   *queryRope,
  const aclTensor   *keyIn,
  const aclTensor   *keyInRope,
  const aclTensor   *value,
  const aclTensor   *dy,
  const aclTensor   *pseShiftOptional,
  const aclTensor   *dropMaskOptional,
  const aclTensor   *paddingMaskOptional,
  const aclTensor   *attenMaskOptional,
  const aclTensor   *softmaxMaxOptional,
  const aclTensor   *softmaxSumOptional,
  const aclTensor   *softmaxInOptional,
  const aclTensor   *attentionInOptional,
  const aclTensor   *sinkInOptional,
  const aclIntArray *prefixOptional,
  const aclIntArray *actualSeqQLenOptional,
  const aclIntArray *actualSeqKvLenOptional,
  const aclIntArray *qStartIdxOptional,
  const aclIntArray *kvStartIdxOptional,
  double             scaleValue,
  double             keepProb,
  int64_t            preTokens,
  int64_t            nextTokens,
  int64_t            headNum,
  char              *inputLayout,
  int64_t            innerPrecise,
  int64_t            sparseMode,
  int64_t            pseType,
  char              *softmaxInLayout,
  const aclTensor   *dqOut,
  const aclTensor   *dqRopeOut,
  const aclTensor   *dkOut,
  const aclTensor   *dkRopeOut,
  const aclTensor   *dvOut,
  const aclTensor   *dpseOut,
  const aclTensor   *dsinkOut,
  uint64_t          *workspaceSize,
  aclOpExecutor    **executor);

aclnnStatus aclnnFlashAttentionUnpaddingScoreGradV5(
  void             *workspace,
  uint64_t          workspaceSize,
  aclOpExecutor    *executor,
  const aclrtStream stream)

aclnnFlashAttentionUnpaddingScoreGradV5GetWorkspaceSize

• Parameters

  Parameter	Input/Output	Description	Instruction	Data Type	Data Format	Dimension (shape)	Non-contiguous Tensor
  query	Input	Q in the formula.	The data type must be consistent with keyIn/value.	FLOAT16、BFLOAT16	ND	[TND]	√
  queryRope	Optional input	The rope part (rotary position encoding) of Q.	Data type must be consistent with query.	BFLOAT16	ND	[TND]	√
  keyIn	Input	K in the formula.	The data type must be consistent with query/value.	FLOAT16, BFLOAT16	ND	[TND]	√
  keyInRope	Optional input	The RoPE part of K.	The data type must be consistent with keyIn.	BFLOAT16	ND	[TND]	√
  value	Input	V in the formula.	The data type must be consistent with that of query/keyIn.	FLOAT16, BFLOAT16	ND	[TND]	√
  dy	Input	dY in the formula.	-	FLOAT16, BFLOAT16	ND	[TND]	√
  pseShiftOptional	Optional input	pse in the formula.	The data type must be consistent with that of query. This parameter must be used together with pseType.	FLOAT16、BFLOAT16	ND	[B,N,1024,Skv]、[1,N,1024,Skv]、[B,N]、[N]	√
  dropMaskOptional	Optional input	Dropout mask.	Incompatible with rope. Pass nullptr if not used.	UINT8	ND	0, 1	√
  paddingMaskOptional	Optional input.	Reserved parameter. Not currently used.	Pass null when calling.	-	-	-	-
  qStartIdxOptional	Optional input	Represents the start index of the current chunk's query sequence in the global context for the external slicing scenario.	-	INT64	ND	0, 1	-
  kvStartIdxOptional	Optional input	Represents the external concatenation scenario, indicating the starting index of the current chunk's key and value sequences in the global sequence.	-	INT64	ND	0, 1	-
  attenMaskOptional	Optional input	atten_mask in the formula.	A value of 1 indicates that the bit does not participate in the computation, and a value of 0 indicates that the bit participates in the computation.	BOOL, UINT8	ND	[B,N,Sq,Skv], [B,1,Sq,Skv], [1,1,Sq,Skv], [Sq,Skv]	√
  softmaxMaxOptional	Optional input	Intermediate output of the forward softmax.	-	FLOAT	ND	[TN8], [NT8]	√
  softmaxSumOptional	Optional input	Intermediate output of the forward softmax.	-	FLOAT	ND	[TN8], [NT8]	√
  softmaxInOptional	Optional input	Intermediate output of the forward softmax.	Reserved parameter, not currently used.	-	-	-	-
  attentionInOptional	Optional input	Forward attention output.	Must be consistent with the data type and shape of query.	FLOAT16, BFLOAT16, FLOAT32	ND	[TND]	√
  sinkInOptional	Optional input.	sink in the formula.	Length is headNum.	FLOAT32	ND	[headNum]	√
  prefixOptional	Optional input.	N of each batch in the prefix sparse scenario.	Pass nullptr if not used.	INT64	ND	0, 1	-
  actualSeqQLenOptional	Optional input	Actual Query sequence length.	-	INT64	ND	1	-
  actualSeqKvLenOptional	Optional input.	Actual Key/Value sequence length.	-	INT64	ND	1	-
  scaleValue	Optional input	Scale scaling factor.	Generally set to D^-0.5.	DOUBLE	-	-	-
  keepProb	Optional input	Proportion of 1s in dropMaskOptional.	Generally set to 1.0.	DOUBLE	-	-	-
  preTokens	Optional input	Left boundary of the sliding window during sparse computation.	2147483647 is recommended if not explicitly specified.	INT64	-	-	-
  nextTokens	Optional input	Right boundary of the sliding window during sparse computation.	2147483647 is recommended if not explicitly specified.	INT64	-	-	-
  headNum	Input	Number of heads on a single card, i.e., the length of the N-axis of Query.	For comprehensive constraints, see Constraints.	INT64	-	-	-
  inputLayout	Input	Input Q/K/V data layout.	Supports TND.	String	-	-	-
  innerPrecise	Optional input.	Internal computation precision control.	Reserved parameter, not currently used.	INT64	-	-	-
  sparseMode	Optional Input	Sparse mode.	Supports configuration of 0 to 8, excluding 5. When rope is passed in, 6 is not supported.	INT64	-	-	-
  pseType	Optional input.	pse type.	Supports configuration values from 0 to 3.	INT64	-	-	-
  softmaxInLayout	Optional input	Controls the actual data layout of softmaxMax and softmaxSum.	Pass "same_as_input" for TND, or "" for NTD.	String	-	-	-
  dqOut	Output	dQ in the formula, the gradient of Query.	-	FLOAT16、BFLOAT16	ND	[TND]	√
  dkOut	Output	dK in the formula, the gradient of Key.	-	FLOAT16, BFLOAT16	ND	[TND]	√
  dvOut	Output	dV in the formula, representing the Value gradient.	-	FLOAT16, BFLOAT16	ND	[TND]	√
  dpseOut	Output	Gradient of d(pse).	Reserved parameter. Not currently used.	-	-	-	-
  dsinkOut	Output	dSink in the formula, representing the gradient of d(sinkInOptional).	-	FLOAT32	ND	[headNum]	√
  workspaceSize	Output	Returns the workspace size to be allocated on the Device side.	-	-	-	-	-
  executor	Output	Returns the operator executor, which contains the operator computation flow.	-	-	-	-	-

• Return Value

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

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

  Return Value	Error Code	Description
  ACLNN_ERR_PARAM_NULLPTR	161001	The input parameter is a required input, output, or required attribute, and is a null pointer.
  ACLNN_ERR_PARAM_INVALID	161002	The data type of query, keyIn, value, dy, pseShiftOptional, dropMaskOptional, paddingMaskOptional, attenMaskOptional, softmaxMaxOptional, softmaxSumOptional, softmaxInOptional, attentionInOptional, sinkInOptional, dqOut, dkOut, dvOut, dsinkOut, or softmaxInLayout is not within the supported range.
  		The data format of query, keyIn, value, dy, pseShiftOptional, dropMaskOptional, paddingMaskOptional, attenMaskOptional, softmaxMaxOptional, softmaxSumOptional, softmaxInOptional, attentionInOptional, sinkInOptional, dqOut, dkOut, dvOut, dsinkOut, or softmaxInLayout is not within the supported range.

aclnnFlashAttentionUnpaddingScoreGradV5

• Parameters

  Parameter	Input/Output	Description
  workspace	Input	Workspace memory address applied for on the Device side.
  workspaceSize	Input	Workspace size applied for on the Device side, obtained from the first-phase API aclnnFlashAttentionUnpaddingScoreGradV5GetWorkspaceSize.
  executor	Input	Operator executor, which contains the operator computation flow.
  stream	Input	Specifies the stream for task execution.

• Return Value

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

Constraints

• Deterministic computation:

  • aclnnFlashAttentionUnpaddingScoreGradV5 defaults to a non-deterministic implementation. Deterministic computation can be enabled by calling aclrtCtxSetSysParamOpt.

• When this API is used with PyTorch, ensure that the CANN-related packages and PyTorch-related packages have matching versions.

• The B (batchsize) dimension of the inputs query, queryRope, key, keyRope, value, and dy must be equal.

• The inputLayout of query, key, value, and dy must be consistent.

• The inputLayout of query, queryRope, key, keyRope, value, and dy must be TND.

• The data type of query, key, value, and pseShiftOptional must be consistent.

• The shape of key and value must be consistent. When the d dimensions of query, key, and value are the same, the shape of query and dy must be consistent. Note: The current version only supports the case where the d dimensions of query, key, and value are the same.

• The N of query and the N of key/value may differ, but they must be proportional, that is, Nq/Nkv must be a non-zero integer, with Nq in the value range of 1 to 256.

• Constraints on data shape, taking TND of inputLayout as an example, where:

  • T(B*S): The value range is 1 to 1M.

  • B: The value range is 1 to 20000. When prefixOptional is present, B supports a maximum of 1K.

  • N: The value range is 1 to 256.

  • S: The value range is 1 to 1M.

  • D: value range: 1 to 768.

  • KeepProb: value range: (0, 1].

• The data layout of query, key, and value supports only TND, where T is the tightly packed data of B and S axes (SeqLenQ and SeqLenKV of each batch), B (Batch) indicates the input sample batch size, S (Seq-Length) indicates the input sample sequence length, H (Head-Size) indicates the hidden layer size, N (Head-Num) indicates the number of heads, and D (Head-Dim) indicates the minimum unit size of the hidden layer, satisfying D = H/N.

• pseShiftOptional: If Sq is greater than 1024, and Sq and Skv of each batch are of equal length, and the scenario is a lower triangular mask with sparseMode 0, 2, or 3, alibi position encoding compression can be enabled. In this case, only the last 1024 rows of the original PSE need to be input, achieving memory optimization, i.e., alibi_compress = ori_pse[:, :, -1024:, :]. Details are as follows:

  • When the parameter differs for each batch, the shape is BNHSkv (H = 1024).

  • When each batch is the same, the shape is 1NHSkv (H=1024).

  • If pseType is 2 or 3, the data type must be FLOAT32, and the corresponding shape supports the range [B,N] or [N].

  • If this parameter is not enabled, pseShiftOptional must be passed as nullptr, and pseType must be passed as 1.

• Meaning of each pseType value

  pseType	Meaning	Note
  0	Externally passed pse, first mul then add.	-
  1	Externally passed pse, first add then mul.	Consistent with the implementation of FlashAttentionScoreGrad.
  2	Internally generated pse, first mul then add.	-
  3	Internally generated pse, first mul then add then sqrt.	-

• The constraints on sparseMode are as follows:

  • When the shapes of all attenMaskOptional are smaller than 2048 and identical, it is recommended to use the default mode to reduce memory usage.

  • When configured as 1, 2, or 3, the user-configured preTokens and nextTokens do not take effect.

  • When sparseMode is configured as 0 or 4, ensure that attenMaskOptional is consistent with the ranges of preTokens and nextTokens.

  • If the user does not specify a value explicitly, it is recommended to pass 0.

  • For details about different sparse modes, see sparse mode description.

  • When configured as 7, the optional input pseShiftOptional is not supported.

  • When configured as 8, the optional input pseShiftOptional is supported when q and kv of each sequence have equal lengths, and pse is generated globally. Outward slicing in the q direction is supported, which requires that q and kv of each sequence have equal lengths before slicing, and after slicing, actualSeqQLenOptional[0] - actualSeqKvLenOptional[0] + qStartIdxOptional - kvStartIdxOptional == 0 (this is an experimental feature).

• For details about different data formats, see data format.

• In some scenarios, if the computation workload is too large, the operator execution may time out (aicore error, with errorStr: timeout or trap error). In this case, axis splitting is recommended. Note: The computation workload here is affected by parameters such as B, S, N, and D. Larger values result in a heavier computation workload.

• The prefixOptional sparse computation only supports the compressed scenario, sparseMode=6. When Sq > Skv, the N value range of prefix is [0, Skv]; when Sq <= Skv, the N value range of prefix is [Skv-Sq, Skv].

• The actualSeqQLenOptional input supports a case where the S length of a batch is 0, but the optional input pseShiftOptional is not supported in this case.

• Constraints on the softmaxMax and softmaxSum parameters: The input format is fixed as [B, N, S, 8], except for the TND input format, which is [T, N, 8]. Note: T=B*S.

• The value of headNum must be consistent with the N value in the input query.

• When the data layout of softmaxSum and softmaxMax is TND, softmaxInLayout must be "same_as_input".

• The dimension of sinkInOptional is 1, and its length must be the same as the headNum of query.

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_flash_attention_score_grad.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;
}

void PrintOutResult(std::vector<int64_t> &shape, void** deviceAddr) {
  auto size = GetShapeSize(shape);
  std::vector<float> resultData(size, 0);
  auto ret = aclrtMemcpy(resultData.data(), resultData.size() * sizeof(resultData[0]),
                           *deviceAddr, 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);
  for (int64_t i = 0; i < size; i++) {
    LOG_PRINT("mean result[%ld] is: %f\n", i, resultData[i]);
  }
}

int Init(int32_t deviceId, aclrtContext* context, 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 = aclrtCreateContext(context, deviceId);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtCreateContext failed. ERROR: %d\n", ret); return ret);
  ret = aclrtSetCurrentContext(*context);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclrtSetCurrentContext 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 apply for device-side memory.
  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 host-side data to device-side 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);

  // Calculate 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.
  // Fill in the deviceId based on your actual device.
  int32_t deviceId = 0;
  aclrtContext context;
  aclrtStream stream;
  auto ret = Init(deviceId, &context, &stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("Init acl failed. ERROR: %d\n", ret); return ret);

  // 2. Construct the inputs and outputs. Custom construction is required based on the API.
  std::vector<int64_t> qShape = {256, 1, 128};
  std::vector<int64_t> kShape = {256, 1, 128};
  std::vector<int64_t> vShape = {256, 1, 128};
  std::vector<int64_t> dxShape = {256, 1, 128};
  std::vector<int64_t> attenmaskShape = {256, 256};
  std::vector<int64_t> softmaxMaxShape = {256, 1, 8};
  std::vector<int64_t> softmaxSumShape = {256, 1, 8};
  std::vector<int64_t> attentionInShape = {256, 1, 128};
  std::vector<int64_t> sinkInOptionalShape = {1};
  std::vector<int64_t> dqShape = {256, 1, 128};
  std::vector<int64_t> dkShape = {256, 1, 128};
  std::vector<int64_t> dvShape = {256, 1, 128};
  std::vector<int64_t> dsinkShape = {1};

  void* qDeviceAddr = nullptr;
  void* kDeviceAddr = nullptr;
  void* vDeviceAddr = nullptr;
  void* dxDeviceAddr = nullptr;
  void* attenmaskDeviceAddr = nullptr;
  void* softmaxMaxDeviceAddr = nullptr;
  void* softmaxSumDeviceAddr = nullptr;
  void* attentionInDeviceAddr = nullptr;
  void* sinkInOptionalDeviceAddr = nullptr;

  void* dqDeviceAddr = nullptr;
  void* dkDeviceAddr = nullptr;
  void* dvDeviceAddr = nullptr;
  void* dsinkDeviceAddr = nullptr;

  aclTensor* q = nullptr;
  aclTensor* k = nullptr;
  aclTensor* v = nullptr;
  aclTensor* dx = nullptr;
  aclTensor* pse = nullptr;
  aclTensor* dropMask = nullptr;
  aclTensor* padding = nullptr;
  aclTensor* attenmask = nullptr;
  aclTensor* softmaxMax = nullptr;
  aclTensor* softmaxSum = nullptr;
  aclTensor* softmaxIn = nullptr;
  aclTensor* attentionIn = nullptr;
  aclTensor* sinkInOptional = nullptr;
  aclTensor* dq = nullptr;
  aclTensor* dk = nullptr;
  aclTensor* dv = nullptr;
  aclTensor* dpse = nullptr;
  aclTensor* dsink = nullptr;

  std::vector<float> qHostData(32768, 1);
  std::vector<float> kHostData(32768, 1);
  std::vector<float> vHostData(32768, 1);
  std::vector<float> dxHostData(32768, 1);
  std::vector<uint8_t> attenmaskHostData(65536, 0);
  std::vector<float> softmaxMaxHostData(2048, 3.0);
  std::vector<float> softmaxSumHostData(2048, 3.0);
  std::vector<float> attentionInHostData(32768, 1);
  std::vector<float> sinkInOptionalHostData(1, 0);
  std::vector<float> dqHostData(32768, 0);
  std::vector<float> dkHostData(32768, 0);
  std::vector<float> dvHostData(32768, 0);
  std::vector<float> dsinkHostData(1, 0);
  ret = CreateAclTensor(qHostData, qShape, &qDeviceAddr, aclDataType::ACL_FLOAT, &q);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(kHostData, kShape, &kDeviceAddr, aclDataType::ACL_FLOAT, &k);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(vHostData, vShape, &vDeviceAddr, aclDataType::ACL_FLOAT, &v);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(dxHostData, dxShape, &dxDeviceAddr, aclDataType::ACL_FLOAT, &dx);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(attenmaskHostData, attenmaskShape, &attenmaskDeviceAddr, aclDataType::ACL_UINT8, &attenmask);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(softmaxMaxHostData, softmaxMaxShape, &softmaxMaxDeviceAddr, aclDataType::ACL_FLOAT, &softmaxMax);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(softmaxSumHostData, softmaxSumShape, &softmaxSumDeviceAddr, aclDataType::ACL_FLOAT, &softmaxSum);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(attentionInHostData, attentionInShape, &attentionInDeviceAddr, aclDataType::ACL_FLOAT, &attentionIn);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(sinkInOptionalHostData, sinkInOptionalShape, &sinkInOptionalDeviceAddr, aclDataType::ACL_FLOAT, &sinkInOptional);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(dqHostData, dqShape, &dqDeviceAddr, aclDataType::ACL_FLOAT, &dq);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(dkHostData, dkShape, &dkDeviceAddr, aclDataType::ACL_FLOAT, &dk);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(dvHostData, dvShape, &dvDeviceAddr, aclDataType::ACL_FLOAT, &dv);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  ret = CreateAclTensor(dsinkHostData, dsinkShape, &dsinkDeviceAddr, aclDataType::ACL_FLOAT, &dsink);
  CHECK_RET(ret == ACL_SUCCESS, return ret);
  std::vector<int64_t> prefixOp = {0};
  aclIntArray* prefix = aclCreateIntArray(prefixOp.data(), 1);
  std::vector<int64_t>  acSeqQLenOp = {256};
  std::vector<int64_t>  acSeqKvLenOp = {256};
  aclIntArray* acSeqQLen = aclCreateIntArray(acSeqQLenOp.data(), acSeqQLenOp.size());
  aclIntArray* acSeqKvLen = aclCreateIntArray(acSeqKvLenOp.data(), acSeqKvLenOp.size());
  std::vector<int64_t> qStartIdxOp = {0};
  std::vector<int64_t> kvStartIdxOp = {0};
  aclIntArray *qStartIdx = aclCreateIntArray(qStartIdxOp.data(), 1);
  aclIntArray *kvStartIdx = aclCreateIntArray(kvStartIdxOp.data(), 1);
  double scaleValue = 0.088388;
  double keepProb = 1;
  int64_t preTokens = 65536;
  int64_t nextTokens = 65536;
  int64_t headNum = 1;
  int64_t innerPrecise = 0;
  int64_t sparseMode = 0;
  int64_t pseType = 1;
  char softmaxInLayoutArr[] = "same_as_input";
  char layOut[5] = {'T', 'N', 'D', 0};

  // 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 aclnnFlashAttentionUnpaddingScoreGradV5.
  ret = aclnnFlashAttentionUnpaddingScoreGradV5GetWorkspaceSize(q, nullptr, k, nullptr, v, dx, pse, dropMask, padding,
              attenmask, softmaxMax, softmaxSum, softmaxIn, attentionIn, sinkInOptional, prefix, acSeqQLen, acSeqKvLen,
              qStartIdx, kvStartIdx, scaleValue, keepProb, preTokens, nextTokens, headNum, layOut, innerPrecise, sparseMode,
              pseType, softmaxInLayoutArr, dq, nullptr, dk, nullptr, dv, dpse, dsink, &workspaceSize, &executor);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnFlashAttentionUnpaddingScoreGradV5GetWorkspaceSize failed. ERROR: %d\n", ret); return ret);

  // Allocate device memory based on the workspaceSize computed 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 aclnnFlashAttentionUnpaddingScoreGradV5.
  ret = aclnnFlashAttentionUnpaddingScoreGradV5(workspaceAddr, workspaceSize, executor, stream);
  CHECK_RET(ret == ACL_SUCCESS, LOG_PRINT("aclnnFlashAttentionUnpaddingScoreGradV5 failed. ERROR: %d\n", ret); return ret);

  // 4. (Boilerplate) Synchronize and 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 side memory to the host side. Modify based on the specific API definition.
  PrintOutResult(dqShape, &dqDeviceAddr);
  PrintOutResult(dkShape, &dkDeviceAddr);
  PrintOutResult(dvShape, &dvDeviceAddr);
  PrintOutResult(dsinkShape, &dsinkDeviceAddr);

  // 6. Release aclTensor and aclScalar. Modify based on the specific API definition.
  aclDestroyTensor(q);
  aclDestroyTensor(k);
  aclDestroyTensor(v);
  aclDestroyTensor(dx);
  aclDestroyTensor(attenmask);
  aclDestroyTensor(softmaxMax);
  aclDestroyTensor(softmaxSum);
  aclDestroyTensor(attentionIn);
  aclDestroyTensor(sinkInOptional);
  aclDestroyTensor(dq);
  aclDestroyTensor(dk);
  aclDestroyTensor(dv);
  aclDestroyTensor(dsink);
  // 7. Release device resources.
  aclrtFree(qDeviceAddr);
  aclrtFree(kDeviceAddr);
  aclrtFree(vDeviceAddr);
  aclrtFree(dxDeviceAddr);
  aclrtFree(attenmaskDeviceAddr);
  aclrtFree(softmaxMaxDeviceAddr);
  aclrtFree(softmaxSumDeviceAddr);
  aclrtFree(attentionInDeviceAddr);
  aclrtFree(sinkInOptionalDeviceAddr);
  aclrtFree(dqDeviceAddr);
  aclrtFree(dkDeviceAddr);
  aclrtFree(dvDeviceAddr);
  aclrtFree(dsinkDeviceAddr);
  if (workspaceSize > 0) {
    aclrtFree(workspaceAddr);
  }
  aclrtDestroyStream(stream);
  aclrtDestroyContext(context);
  aclrtResetDevice(deviceId);
  aclFinalize();

  return 0;
}