Developer Reference for Intel® oneAPI Math Kernel Library for Fortran

ID 766686
Date 3/31/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

gemm_*_pack

Pack the matrix into the buffer allocated previously.

Syntax

call gemm_s8u8s32_pack (identifier, trans, m, n, k, src, ld, dest)

call gemm_s16s16s32_pack (identifier, trans, m, n, k, src, ld, dest)

Include Files
  • mkl.fi
Description

The gemm_*_pack routine is one of a set of related routines that enable the use of an internal packed storage. Call gemm_*_pack after you allocate a buffer whose size is given by gemm_*_pack_get_size. The gemm_*_pack routine packs the identified matrix into the buffer allocated previously.

The gemm_*_pack routine performs this operation:

dest := op(src) as part of the computation C := alpha*(op(A) + A_offset)*(op(B) + B_offset) + beta*C + C_offset for integer types.

C := alpha*op(A) * op(B) + beta*C for bfloat16 type.

where:

  • op(X) is one of the operations op(X) = X or op(X) = XT
  • alpha and beta are scalars,
  • src is a matrix,
  • A , A_offset,B, B_offset,c,and C_offset are matrices
  • op(src) is an m-by-k matrix if identifier = 'A' or 'a',
  • op(src) is a k-by-n matrix if identifier = 'B' or 'b',
  • dest is the buffer previously allocated to store the matrix packed into an internal format
  • A_offset is an m-by-k matrix.
  • B_offset is an k-by-n matrix.
  • C_offset is an m-by-n matrix.

NOTE:

For best performance, use the same number of threads for packing and for computing.

If packing for both A and B matrices, you must use the same number of threads for packing A as for packing B.

Input Parameters
identifier
CHARACTER*1.

Specifies which matrix is to be packed:

If identifier = 'A' or 'a', the A matrix is packed.

If identifier = 'B' or 'b', the B matrix is packed.

trans
CHARACTER*1.

Specifies the form of op(src) used in the packing:

If trans = 'N' or 'n'  op(src) = src.

If trans = 'T' or 't'  op(src) = srcT.

m
INTEGER.

Specifies the number of rows of matrix op(A) and of the matrix C. The value of m must be at least zero.

n
INTEGER.

Specifies the number of columns of matrix op(B) and the number of columns of matrix C. The value of n must be at least zero.

k
INTEGER.

Specifies the number of columns of matrix op(A) and the number of rows of matrix op(B). The value of k must be at least zero.

src

INTEGER*1 for gemm_s8u8s32_pack and INTEGER*2 for gemm_s16s16s32_pack

 

trans = 'N' or 'n'

trans = 'T' or 't'

identifier = 'A' or 'a'

Size ld*k.

Before entry, the leading m-by-k part of the array src must contain the matrix A.

Size ld*m.

Before entry, the leading k-by-m part of the array src must contain the matrix A.

identifier = 'B' or 'b'

Size ld*n.

Before entry, the leading k-by-n part of the array src must contain the matrix B.

Size ld*k.

Before entry, the leading n-by-k part of the array src must contain the matrix B.

ld

INTEGER. Specifies the leading dimension of src as declared in the calling (sub)program.

 

trans = 'N' or 'n'

trans = 'T' or 't'

identifier = 'A' or 'a'

ld must be at least max(1, m).

ld must be at least max(1, k).

identifier = 'B' or 'b'

ld must be at least max(1, k).

ld must be at least max(1, n).

dest
INTEGER*1 for gemm_s8u8s32_pack or INTEGER*2 for gemm_s16s16s32_pack

Buffer for the packed matrix.

Output Parameters

dest

INTEGER*1 for gemm_s8u8s32_pack or INTEGER*2 for gemm_s16s16s32_pack

Overwritten by the matrix op(src)stored in a format internal to Intel® oneAPI Math Kernel Library.

Example

See the following examples in the MKL installation directory to understand the use of these routines:

gemm_s8u8s32_pack: examples\blas\source\gemm_s8u8s32_computex.f

gemm_s16s16s32_pack: examples\blas\source\gemm_s16s16s32_computex.f

Application Notes

See Also