Cgemm

Applicable Products

Function Description

API function:

• asdBlasMakeCgemmPlan: initializes the Cgemm operator configuration corresponding to the handle.
• asdBlasCgemm: a type of matrix multiplication operation, which is used to compute the product of two complex matrices.

Formula:

• asdBlasCgemm formula:

  where,

• Example:

  The inTensorA is as follows:

  [   [ 1+i, 1+2i ],
      [ 1+3i, 1+4i ]  ]

  The inTensorB is as follows:

  [   [ 2+i, 2+2i ],
      [ 2+3i, 2+4i ]  ]

  The input inTensorC is as follows:

  [   [ 3+i, 3+2i ],
      [ 3+3i, 3+4i ]  ]

  The input transa is N, and the input transb is T.

  The input m is 2, the input n is 2, the input k is 2, the input alpha is 1+i, and the input beta is 2+2i.

  The input lda is 2, the input ldb is 2, and the input ldc is 2.

  After the Cgemm operator is called, the outTensor is as follows:

  [   [ -15+19i, -27+19i ],
      [ -37+21i, -57+13i ]  ]

Function Prototype

• AspbStatus asdBlasMakeCgemmPlan(asdBlasHandle handle, asdBlasOperation_t transa, asdBlasOperation_t transb, int64_t m,int64_t n, int64_t k, int64_t lda, int64_t ldb, int64_t ldc)
• AspbStatus asdBlasCgemm(asdBlasHandle handle, asdBlasOperation_t transa, asdBlasOperation_t transb, const int64_t m,const int64_t n, const int64_t k, const std::complex<float> *alpha, aclTensor *A,const int64_t lda, aclTensor *B, const int64_t ldb, const std::complex<float> *beta,aclTensor *C, const int64_t ldc)

Parameter Description

• asdBlasMakeCgemmPlan

  Parameter	Input/Output	Type	Description
  handle	Input	asdBlasHandle	Handle of the Cgemm operator.
  transa	Input	asdBlasOperation_t	Whether matrix A should be transposed. ASDBLAS_OP_N //Not transposed ASDBLAS_OP_T // Transposed ASDBLAS_OP_C //Conjugately transposed
  transb	Input	asdBlasOperation_t	Whether matrix B needs to be transposed. ASDBLAS_OP_N //Not transposed ASDBLAS_OP_T // Transposed ASDBLAS_OP_C //Conjugately transposed
  m	Input	const int64_t	Number of rows in matrix A and matrix C.
  n	Input	const int64_t	Number of columns in matrix B and matrix C.
  k	Input	const int64_t	Common dimension of matrix A and matrix B.
  lda	Input	const int64_t	Memory address offset between adjacent elements in matrix A (currently constrained to m).
  ldb	Input	const int64_t	Memory address offset between adjacent elements in matrix B (currently constrained to k).
  ldc	Input	const int64_t	Memory address offset between adjacent elements in matrix C (currently constrained to m).

• asdBlasCgemm

  Parameter	Input/Output	Type	Description
  handle	Input	asdBlasHandle	cgemm operator handle.
  transa	Input	asdBlasOperation_t	Whether matrix A should be transposed. ASDBLAS_OP_N // Not transposed ASDBLAS_OP_T // Transposed ASDBLAS_OP_C // Conjugately transposed
  transb	Input	asdBlasOperation_t	Whether matrix B needs to be transposed. ASDBLAS_OP_N // Not transposed ASDBLAS_OP_T // Transposed ASDBLAS_OP_C // Conjugately transposed
  m	Input	const int64_t	Number of rows in matrix C.
  n	Input	const int64_t	Number of columns in matrix C.
  k	Input	const int64_t	Common dimension of matrix A and matrix B.
  alpha	Input	const std::complex<float> *	alpha in the formula, which is a complex scalar used to multiply the result of matrix multiplication.
  A	Input	aclTensor *	A in the formula, which is in column-major order and is a tensor on the device. The data type is COMPLEX64, the data format is ND, and the shape is [m, k].
  lda	Input	const int64_t	Memory address offset between adjacent elements in matrix A (currently constrained to m).
  B	Input	aclTensor *	B in the formula, which is a tensor on the device. The data type can only be COMPLEX64, the data format can be ND, and the shape is [k, n].
  ldb	Input	const int64_t	Memory address offset between adjacent elements in matrix B (currently constrained to k).
  beta	Input	const std::complex<float> *	beta in the formula is a complex scalar, and is used to multiply the matrix C.
  C	Input/Output	aclTensor *	C in the formula, which is a tensor on the device. The data type can only be COMPLEX64, the data format can be ND, and the shape is [m, n].
  ldc	Input	const int64_t	Memory address offset between adjacent elements in matrix C (currently constrained to m).

Return Value Description

For details about the return values, see Return Value.

Constraints

• asdBlasMakeCgemmPlan: none.
• asdBlasCgemm
  • The input element count (m, n, k) currently supports the range [1, 8192].
  • The operator input data is in column-major order. The input shape is [m, k], [k, n], or [m, n], and the output shape is [m, n].
  • The operator does not support the calculation for three or more dimensions.

Calling Example

For details about the operator calling example, see Cgemm.