MPCRTileGemm: MPCRTile GEMM (Matrix-Matrix Multiplication)

In MPCR: Multi- And Mixed-Precision Computations

MPCRTile GEMM (Matrix-Matrix Multiplication)

Description

Tile-based matrix-matrix multiplication of two given MPCR tiled matrices to perform:
C = alpha*A X B + beta*C

Usage

## S4 method for signature 'Rcpp_MPCRTile'
MPCRTile.gemm(a,b,c,transpose_a= FALSE,transpose_b=FALSE,alpha=1,beta=0,num_threads=1)
   

Arguments

a	An MPCR tile matrix A.
b	An MPCR tile matrix B.
c	Input/Output MPCR tile matrix C.
transpose_a	A flag to indicate whether transpose matrix A should be used.
transpose_b	A flag to indicate whether transpose matrix B should be used.
alpha	Specifies the scalar alpha.
beta	Specifies the scalar beta.
num_threads	An integer to determine number if thread to run using openmp, default = 1 (serial with no parallelization).

Value

An MPCR tile matrix C.

MPCR documentation built on April 13, 2025, 5:08 p.m.