?hgemm

Compute the product of a general matrix and a matrix.

That is, $\text{[math]}$ .

The value of op(X) may be $\text{[math]}$ . alpha and beta are multiplication coefficients; op(A) is an m*k matrix; op(B) is a k*n matrix, and C is an m*n matrix.

Interface Definition

C interface:

void cblas_hgemm(const enum CBLAS_ORDER order, const enum CBLAS_TRANSPOSE transA, const enum CBLAS_TRANSPOSE transB, const BLASINT M, const BLASINT N, const BLASINT K, const __fp16 alpha, const __fp16 *A, const BLASINT lda, const __fp16 *B, const BLASINT ldb, const __fp16 beta, __fp16 *C, const BLASINT ldc);

void cblas_shgemm(const enum CBLAS_ORDER order, const enum CBLAS_TRANSPOSE transA, const enum CBLAS_TRANSPOSE transB, const BLASINT M, const BLASINT N, const BLASINT K, const float alpha, const __fp16 *A, const BLASINT lda, const __fp16 *B, const BLASINT ldb, const float beta, float *C, const BLASINT ldc);

void cblas_chgemm(const enum CBLAS_ORDER order, const enum CBLAS_TRANSPOSE transA, const enum CBLAS_TRANSPOSE transB, const BLASINT M, const BLASINT N, const BLASINT K, const void *alpha, const void *A, const BLASINT lda, const void *B, const BLASINT ldb, const void *beta, void *C, const BLASINT ldc);

void cblas_cshgemm(const enum CBLAS_ORDER order, const enum CBLAS_TRANSPOSE transA, const enum CBLAS_TRANSPOSE transB, const BLASINT M, const BLASINT N, const BLASINT K, const void *alpha, const void *A, const BLASINT lda, const void *B, const BLASINT ldb, const void *beta, void *C, const BLASINT ldc);

Fortran interface:

CALL HGEMM(TRANSA, TRANSB, M, N, K, ALPHA, A, LDA, B, LDB, BETA, C, LDC)

CALL SHGEMM(TRANSA, TRANSB, M, N, K, ALPHA, A, LDA, B, LDB, BETA, C, LDC)

CALL CHGEMM(TRANSA, TRANSB, M, N, K, ALPHA, A, LDA, B, LDB, BETA, C, LDC)

CALL CSHGEMM(TRANSA, TRANSB, M, N, K, ALPHA, A, LDA, B, LDB, BETA, C, LDC)

Parameters

Parameter	Type	Description	Input/Output
order	Enumeration type CBLAS_ORDER	Whether the matrix is in row- or column-major order.	Input
TransA	Enumeration type CBLAS_TRANSPOSE	Whether matrix A is non-transposed, transposed, or conjugated. If TransA = CblasNoTrans, . If TransA = CblasTrans, . If TransA = CblasConjTrans, $\text{[math]}$ . If TransA = CblasConjTrans, .	Input
TransB	Enumeration type CBLAS_TRANSPOSE	Whether matrix B is non-transposed, transposed, or conjugated. If TransB = CblasNoTrans, . If TransB = CblasTrans, . If TransB = CblasConjTrans, $\text{[math]}$ . If TransB = CblasConjTrans, .	Input
M	Integer	Number of rows in matrices op(A) and C.	Input
N	Integer	Number of columns in matrices op(B) and C.	Input
K	Integer	Number of columns in matrix op(A) and number of rows in matrix op(B).	Input
alpha	Half-precision floating-point type for hgemm Single-precision floating-point type for shgemm Half-precision complex type for chgemm Single-precision complex type for cshgemm	Multiplication coefficient.	Input
A	Half-precision floating-point type for hgemm Half-precision floating-point type for shgemm Half-precision complex type for chgemm Half-precision complex type for cshgemm	Matrix A.	Input
lda	Integer	For a column-store matrix with TransA = CblasNoTrans, lda must be at least max(1, m); otherwise, lda must be at least max(1, k). For a row-store matrix with TransA = CblasNoTrans, lda must be at least max(1, k); otherwise, lda must be at least max(1, m).	Input
B	Half-precision floating-point type for hgemm Half-precision floating-point type for shgemm Half-precision complex type for chgemm Half-precision complex type for cshgemm	Matrix B.	Input
ldb	Integer	For a column-store matrix with TransB = CblasNoTrans, ldb must be at least max(1, k); otherwise, ldb must be at least max(1, n). For a row-store matrix with TransB = CblasNoTrans, ldb must be at least max(1, n); otherwise, ldb must be at least max(1, k).	Input
beta	Half-precision floating-point type for hgemm Single-precision floating-point type for shgemm Half-precision complex type for chgemm Single-precision complex type for cshgemm	Multiplication coefficient.	Input
C	Half-precision floating-point type for hgemm Single-precision floating-point type for shgemm Half-precision complex type for chgemm Single-precision complex type for cshgemm	Matrix C.	Input/Output
ldc	Integer	For a column-store matrix, ldc must be at least max(1, m); otherwise, ldc must be at least max(1, n).	Input

Dependencies

#include "kblas.h"

Examples

C interface:

    int m = 4, k = 3, n = 4, lda = 4, ldb = 3, ldc = 4;  
    __fp16 alpha = 1.0, beta = 2.0;  
     /*  
     * A:  
     *     0.340188,       0.411647,       -0.222225,  
     *     -0.105617,      -0.302449,      0.053970,  
     *     0.283099,       -0.164777,      -0.022603,  
     *     0.298440,       0.268230,       0.128871,  
     * B:  
     *     -0.135216,      0.416195,       -0.358397,      -0.257113,  
     *     0.013401,       0.135712,       0.106969,       -0.362768,  
     *     0.452230,       0.217297,       -0.483699,      0.304177,  
     * C:  
     *     -0.343321,      0.498924,       0.112640,       -0.006417,  
     *     -0.099056,      -0.281743,      -0.203968,      0.472775,  
     *     -0.370210,      0.012932,       0.137552,       -0.207483,  
     *     -0.391191,      0.339112,       0.024287,       0.271358,  
     */  
    __fp16 a[12] = {0.340188, -0.105617, 0.283099,  
                    0.298440, 0.411647, -0.302449,  
                    -0.164777, 0.268230, -0.222225,  
                    0.053970, -0.022603, 0.128871};  
    __fp16 b[12] = {-0.135216, 0.013401, 0.452230, 0.416195,  
                    0.135712, 0.217297, -0.358397, 0.106969,  
                    -0.483699, -0.257113, -0.362768, 0.304177};  
    __fp16 c[16] = {-0.343321, -0.099056, -0.370210, -0.391191,  
                    0.498924, -0.281743, 0.012932, 0.339112,  
                    0.112640, -0.203968, 0.137552, 0.024287,  
                    -0.006417, 0.472775, -0.207483, 0.271358};  
  
    cblas_hgemm(CblasColMajor,CblasNoTrans,CblasNoTrans, m, n, k, alpha, a, lda, b, ldb, beta, c, ldc);  
    /*  
     * Output C:  
     *     -0.827621       1.147010        0.254881        -0.317229  
     *     -0.163476       -0.636762       -0.428542       1.098841  
     *     -0.791128       0.116416        0.166949        -0.434854  
     *     -0.760862       0.866839        -0.092028       0.407877  
     *  
     */

Fortran interface:

      INTEGER :: M=4, K=3, N=4  
      INTEGER :: LDA=4, LDB=3, LDC=4  
      REAL(4) :: ALPHA=1.0, BETA=2.0  
      REAL(4) :: A(12), B(12), C(16)  
      DATA A/0.340188, -0.105617, 0.283099,  
     $       0.298440, 0.411647, -0.302449,  
     $       -0.164777, 0.268230, -0.222225,  
     $       0.053970, -0.022603, 0.128871/  
      DATA B/-0.135216, 0.013401, 0.452230, 0.416195,  
     $       0.135712, 0.217297, -0.358397, 0.106969,  
     $       -0.483699, -0.257113, -0.362768, 0.304177/  
      DATA C/-0.343321, -0.099056, -0.370210, -0.391191,  
     $       0.498924, -0.281743, 0.012932, 0.339112,  
     $       0.112640, -0.203968, 0.137552, 0.024287,  
     $       -0.006417, 0.472775, -0.207483, 0.271358/  
      EXTERNAL HGEMM  
      CALL HGEMM('N', 'N', M, N, K, ALPHA, A, LDA, B, LDB, BETA, C,  
     $          LDC)  
*     Output C:  
*         -0.827621       1.147010        0.254881        -0.317229  
*         -0.163476       -0.636762       -0.428542       1.098841  
*         -0.791128       0.116416        0.166949        -0.434854  
*         -0.760862       0.866839        -0.092028       0.407877

Parent topic: KML_BLAS Level 3 Functions